diff --git a/src/cmd/compile/internal/ssa/_gen/LOONG64.rules b/src/cmd/compile/internal/ssa/_gen/LOONG64.rules
index 74bc60a3027..33cc8a6ae3a 100644
--- a/src/cmd/compile/internal/ssa/_gen/LOONG64.rules
+++ b/src/cmd/compile/internal/ssa/_gen/LOONG64.rules
@@ -42,6 +42,10 @@
 (Select1 <t> (Add64carry x y c)) =>
 	(OR (SGTU <t> x s:(ADDV <t> x y)) (SGTU <t> s (ADDV <t> s c)))
 
+(Select0 <t> (Sub64borrow x y c)) => (SUBV (SUBV <t> x y) c)
+(Select1 <t> (Sub64borrow x y c)) =>
+	(OR (SGTU <t> s:(SUBV <t> x y) x) (SGTU <t> (SUBV <t> s c) s))
+
 // (x + y) / 2 with x>=y => (x - y) / 2 + y
 (Avg64u <t> x y) => (ADDV (SRLVconst <t> (SUBV <t> x y) [1]) y)
 
diff --git a/src/cmd/compile/internal/ssa/rewriteLOONG64.go b/src/cmd/compile/internal/ssa/rewriteLOONG64.go
index e2d0d20efb3..fbe49e58194 100644
--- a/src/cmd/compile/internal/ssa/rewriteLOONG64.go
+++ b/src/cmd/compile/internal/ssa/rewriteLOONG64.go
@@ -6844,6 +6844,22 @@ func rewriteValueLOONG64_OpSelect0(v *Value) bool {
 		v.AddArg2(v0, c)
 		return true
 	}
+	// match: (Select0 <t> (Sub64borrow x y c))
+	// result: (SUBV (SUBV <t> x y) c)
+	for {
+		t := v.Type
+		if v_0.Op != OpSub64borrow {
+			break
+		}
+		c := v_0.Args[2]
+		x := v_0.Args[0]
+		y := v_0.Args[1]
+		v.reset(OpLOONG64SUBV)
+		v0 := b.NewValue0(v.Pos, OpLOONG64SUBV, t)
+		v0.AddArg2(x, y)
+		v.AddArg2(v0, c)
+		return true
+	}
 	// match: (Select0 (DIVVU _ (MOVVconst [1])))
 	// result: (MOVVconst [0])
 	for {
@@ -6978,6 +6994,28 @@ func rewriteValueLOONG64_OpSelect1(v *Value) bool {
 		v.AddArg2(v0, v2)
 		return true
 	}
+	// match: (Select1 <t> (Sub64borrow x y c))
+	// result: (OR (SGTU <t> s:(SUBV <t> x y) x) (SGTU <t> (SUBV <t> s c) s))
+	for {
+		t := v.Type
+		if v_0.Op != OpSub64borrow {
+			break
+		}
+		c := v_0.Args[2]
+		x := v_0.Args[0]
+		y := v_0.Args[1]
+		v.reset(OpLOONG64OR)
+		v0 := b.NewValue0(v.Pos, OpLOONG64SGTU, t)
+		s := b.NewValue0(v.Pos, OpLOONG64SUBV, t)
+		s.AddArg2(x, y)
+		v0.AddArg2(s, x)
+		v2 := b.NewValue0(v.Pos, OpLOONG64SGTU, t)
+		v3 := b.NewValue0(v.Pos, OpLOONG64SUBV, t)
+		v3.AddArg2(s, c)
+		v2.AddArg2(v3, s)
+		v.AddArg2(v0, v2)
+		return true
+	}
 	// match: (Select1 (MULVU x (MOVVconst [-1])))
 	// result: (NEGV x)
 	for {
diff --git a/src/cmd/compile/internal/ssagen/ssa.go b/src/cmd/compile/internal/ssagen/ssa.go
index f4601c56cf7..77307f4c2df 100644
--- a/src/cmd/compile/internal/ssagen/ssa.go
+++ b/src/cmd/compile/internal/ssagen/ssa.go
@@ -4736,8 +4736,8 @@ func InitTables() {
 		func(s *state, n *ir.CallExpr, args []*ssa.Value) *ssa.Value {
 			return s.newValue3(ssa.OpSub64borrow, types.NewTuple(types.Types[types.TUINT64], types.Types[types.TUINT64]), args[0], args[1], args[2])
 		},
-		sys.AMD64, sys.ARM64, sys.PPC64, sys.S390X, sys.RISCV64)
-	alias("math/bits", "Sub", "math/bits", "Sub64", sys.ArchAMD64, sys.ArchARM64, sys.ArchS390X, sys.ArchRISCV64)
+		sys.AMD64, sys.ARM64, sys.PPC64, sys.S390X, sys.RISCV64, sys.Loong64)
+	alias("math/bits", "Sub", "math/bits", "Sub64", p8...)
 	addF("math/bits", "Div64",
 		func(s *state, n *ir.CallExpr, args []*ssa.Value) *ssa.Value {
 			// check for divide-by-zero/overflow and panic with appropriate message
diff --git a/test/codegen/mathbits.go b/test/codegen/mathbits.go
index bd23b90b9b4..acc9930c61d 100644
--- a/test/codegen/mathbits.go
+++ b/test/codegen/mathbits.go
@@ -627,6 +627,7 @@ func Add64MPanicOnOverflowGT(a, b [2]uint64) [2]uint64 {
 func Sub(x, y, ci uint) (r, co uint) {
 	// amd64:"NEGL","SBBQ","NEGQ"
 	// arm64:"NEGS","SBCS","NGC","NEG",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", "SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", "SUBE", "SUBZE", "NEG"
 	// s390x:"SUBE"
@@ -637,6 +638,7 @@ func Sub(x, y, ci uint) (r, co uint) {
 func SubC(x, ci uint) (r, co uint) {
 	// amd64:"NEGL","SBBQ","NEGQ"
 	// arm64:"NEGS","SBCS","NGC","NEG",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", "SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", "SUBE", "SUBZE", "NEG"
 	// s390x:"SUBE"
@@ -647,6 +649,7 @@ func SubC(x, ci uint) (r, co uint) {
 func SubZ(x, y uint) (r, co uint) {
 	// amd64:"SUBQ","SBBQ","NEGQ",-"NEGL"
 	// arm64:"SUBS","NGC","NEG",-"SBCS",-"ADD",-"SUB\t",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", -"SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", -"SUBE", "SUBZE", "NEG"
 	// s390x:"SUBC"
@@ -657,6 +660,7 @@ func SubZ(x, y uint) (r, co uint) {
 func SubR(x, y, ci uint) uint {
 	// amd64:"NEGL","SBBQ",-"NEGQ"
 	// arm64:"NEGS","SBCS",-"NGC",-"NEG\t",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV",-"SGTU"
 	// ppc64:"SUBC", "SUBE", -"SUBZE", -"NEG"
 	// ppc64le:"SUBC", "SUBE", -"SUBZE", -"NEG"
 	// s390x:"SUBE"
@@ -679,6 +683,7 @@ func SubM(p, q, r *[3]uint) {
 func Sub64(x, y, ci uint64) (r, co uint64) {
 	// amd64:"NEGL","SBBQ","NEGQ"
 	// arm64:"NEGS","SBCS","NGC","NEG",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", "SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", "SUBE", "SUBZE", "NEG"
 	// s390x:"SUBE"
@@ -689,6 +694,7 @@ func Sub64(x, y, ci uint64) (r, co uint64) {
 func Sub64C(x, ci uint64) (r, co uint64) {
 	// amd64:"NEGL","SBBQ","NEGQ"
 	// arm64:"NEGS","SBCS","NGC","NEG",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", "SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", "SUBE", "SUBZE", "NEG"
 	// s390x:"SUBE"
@@ -699,6 +705,7 @@ func Sub64C(x, ci uint64) (r, co uint64) {
 func Sub64Z(x, y uint64) (r, co uint64) {
 	// amd64:"SUBQ","SBBQ","NEGQ",-"NEGL"
 	// arm64:"SUBS","NGC","NEG",-"SBCS",-"ADD",-"SUB\t",-"CMP"
+	// loong64:"SUBV","SGTU"
 	// ppc64:"SUBC", -"SUBE", "SUBZE", "NEG"
 	// ppc64le:"SUBC", -"SUBE", "SUBZE", "NEG"
 	// s390x:"SUBC"
@@ -709,6 +716,7 @@ func Sub64Z(x, y uint64) (r, co uint64) {
 func Sub64R(x, y, ci uint64) uint64 {
 	// amd64:"NEGL","SBBQ",-"NEGQ"
 	// arm64:"NEGS","SBCS",-"NGC",-"NEG\t",-"ADD",-"SUB",-"CMP"
+	// loong64:"SUBV",-"SGTU"
 	// ppc64:"SUBC", "SUBE", -"SUBZE", -"NEG"
 	// ppc64le:"SUBC", "SUBE", -"SUBZE", -"NEG"
 	// s390x:"SUBE"