diff --git a/flang/docs/Intrinsics.md b/flang/docs/Intrinsics.md
--- a/flang/docs/Intrinsics.md
+++ b/flang/docs/Intrinsics.md
@@ -657,6 +657,11 @@
 ```
 
 ## Non-standard intrinsics
+### FLANG
+```
+FMADD(REAL F1, REAL F2, REAL F3) -> REAL
+```
+
 ### PGI
 ```
 AND, OR, XOR
diff --git a/flang/lib/Evaluate/intrinsics.cpp b/flang/lib/Evaluate/intrinsics.cpp
--- a/flang/lib/Evaluate/intrinsics.cpp
+++ b/flang/lib/Evaluate/intrinsics.cpp
@@ -495,6 +495,9 @@
             {"back", AnyLogical, Rank::scalar, Optionality::optional}},
         KINDInt, Rank::vector, IntrinsicClass::transformationalFunction},
     {"floor", {{"a", AnyReal}, DefaultingKIND}, KINDInt},
+    {"fmadd",
+        {{"f1", SameFloating}, {"f2", SameFloating}, {"f3", SameFloating}},
+        SameFloating},
     {"fraction", {{"x", SameReal}}, SameReal},
     {"gamma", {{"x", SameReal}}, SameReal},
     {"get_team", {{"level", DefaultInt, Rank::scalar, Optionality::optional}},
diff --git a/flang/lib/Optimizer/Builder/IntrinsicCall.cpp b/flang/lib/Optimizer/Builder/IntrinsicCall.cpp
--- a/flang/lib/Optimizer/Builder/IntrinsicCall.cpp
+++ b/flang/lib/Optimizer/Builder/IntrinsicCall.cpp
@@ -1023,6 +1023,10 @@
     // math::FloorOp returns a real, while Fortran FLOOR returns integer.
     {"floor", "floorf", genF32F32FuncType, genMathOp<mlir::math::FloorOp>},
     {"floor", "floor", genF64F64FuncType, genMathOp<mlir::math::FloorOp>},
+    {"fmadd", "llvm.fma.f32", genF32F32F32F32FuncType,
+     genMathOp<mlir::math::FmaOp>},
+    {"fmadd", "llvm.fma.f64", genF64F64F64F64FuncType,
+     genMathOp<mlir::math::FmaOp>},
     {"gamma", "tgammaf", genF32F32FuncType, genLibCall},
     {"gamma", "tgamma", genF64F64FuncType, genLibCall},
     {"hypot", "hypotf", genF32F32F32FuncType, genLibCall},
diff --git a/flang/test/Lower/Intrinsics/fmadd.f90 b/flang/test/Lower/Intrinsics/fmadd.f90
new file mode 100644
--- /dev/null
+++ b/flang/test/Lower/Intrinsics/fmadd.f90
@@ -0,0 +1,20 @@
+! RUN: bbc -emit-fir %s -o - | FileCheck --check-prefix=CHECK-FIR %s
+! RUN: %flang_fc1 -emit-llvm %s -o - | FileCheck --check-prefix=CHECK-LLVMIR %s
+
+function test_real4(a, x, y)
+  real :: a, x, y
+  test_real4 = fmadd(a, x, y)
+end function
+
+! CHECK-LABEL: @_QPtest_real4
+! CHECK-FIR: {{%[A-Za-z0-9._]+}} = math.fma {{%[0-9]+}}, {{%[0-9]+}}, {{%[0-9]+}} {{.*}} : f32
+! CHECK-LLVMIR: {{%[A-Za-z0-9._]+}} = call {{.*}} float @llvm.fma.f32(float {{%[0-9]+}}, float {{%[0-9]+}}, float {{%[0-9]+}})
+
+function test_real8(a, x, y)
+  real(8) :: a, x, y
+  test_real8 = fmadd(a, x, y)
+end function
+
+! CHECK-LABEL: @_QPtest_real8
+! CHECK-FIR: {{%[A-Za-z0-9._]+}} = math.fma {{%[0-9]+}}, {{%[0-9]+}}, {{%[0-9]+}} {{.*}} : f64
+! CHECK-LLVMIR: {{%[A-Za-z0-9._]+}} = call {{.*}} double @llvm.fma.f64(double {{%[0-9]+}}, double {{%[0-9]+}}, double {{%[0-9]+}})