diff --git a/src/back/hlsl/writer.rs b/src/back/hlsl/writer.rs
index 97a141e42b..3701d0002a 100644
--- a/src/back/hlsl/writer.rs
+++ b/src/back/hlsl/writer.rs
@@ -1795,6 +1795,24 @@ impl<'a, W: fmt::Write> super::Writer<'a, W> {
                 self.write_expr(module, left, func_ctx)?;
                 write!(self.out, ")")?;
             }
+            // While HLSL supports float operands with the % operator it is only
+            // defined in cases where both sides are either positive or negative.
+            Expression::Binary {
+                op: crate::BinaryOperator::Modulo,
+                left,
+                right,
+            } if func_ctx.info[left]
+                .ty
+                .inner_with(&module.types)
+                .scalar_kind()
+                == Some(crate::ScalarKind::Float) =>
+            {
+                write!(self.out, "fmod(")?;
+                self.write_expr(module, left, func_ctx)?;
+                write!(self.out, ", ")?;
+                self.write_expr(module, right, func_ctx)?;
+                write!(self.out, ")")?;
+            }
             Expression::Binary { op, left, right } => {
                 write!(self.out, "(")?;
                 self.write_expr(module, left, func_ctx)?;
diff --git a/tests/out/hlsl/operators.hlsl b/tests/out/hlsl/operators.hlsl
index 6eb3994ac2..e6356d201f 100644
--- a/tests/out/hlsl/operators.hlsl
+++ b/tests/out/hlsl/operators.hlsl
@@ -110,10 +110,10 @@ void arithmetic()
     float4 unnamed_36 = ((2.0).xxxx / (1.0).xxxx);
     int unnamed_37 = (2 % 1);
     uint unnamed_38 = (2u % 1u);
-    float unnamed_39 = (2.0 % 1.0);
+    float unnamed_39 = fmod(2.0, 1.0);
     int2 unnamed_40 = ((2).xx % (1).xx);
     uint3 unnamed_41 = ((2u).xxx % (1u).xxx);
-    float4 unnamed_42 = ((2.0).xxxx % (1.0).xxxx);
+    float4 unnamed_42 = fmod((2.0).xxxx, (1.0).xxxx);
     int2 unnamed_43 = ((2).xx + (1).xx);
     int2 unnamed_44 = ((2).xx + (1).xx);
     uint2 unnamed_45 = ((2u).xx + (1u).xx);
@@ -142,8 +142,8 @@ void arithmetic()
     int2 unnamed_68 = ((2).xx % (1).xx);
     uint2 unnamed_69 = ((2u).xx % (1u).xx);
     uint2 unnamed_70 = ((2u).xx % (1u).xx);
-    float2 unnamed_71 = ((2.0).xx % (1.0).xx);
-    float2 unnamed_72 = ((2.0).xx % (1.0).xx);
+    float2 unnamed_71 = fmod((2.0).xx, (1.0).xx);
+    float2 unnamed_72 = fmod((2.0).xx, (1.0).xx);
     float3x3 unnamed_73 = (float3x3(float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0)) + float3x3(float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0)));
     float3x3 unnamed_74 = (float3x3(float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0)) - float3x3(float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0)));
     float3x3 unnamed_75 = mul(1.0, float3x3(float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0), float3(0.0, 0.0, 0.0)));