diff --git a/cpp/src/gandiva/precompiled/extended_math_ops.cc b/cpp/src/gandiva/precompiled/extended_math_ops.cc
index 3329acc15e49..686d4477ffd5 100644
--- a/cpp/src/gandiva/precompiled/extended_math_ops.cc
+++ b/cpp/src/gandiva/precompiled/extended_math_ops.cc
@@ -116,8 +116,7 @@ POWER(float64, float64, float64)
 #define ROUND_DECIMAL(TYPE)                                                 \
   FORCE_INLINE                                                              \
   gdv_##TYPE round_##TYPE##_int32(gdv_##TYPE number, gdv_int32 out_scale) { \
-    gdv_##TYPE scale_multiplier =                                           \
-        static_cast<gdv_##TYPE>(get_scale_multiplier(out_scale));           \
+    gdv_float64 scale_multiplier = get_scale_multiplier(out_scale);         \
     return static_cast<gdv_##TYPE>(                                         \
         trunc(number * scale_multiplier + ((number > 0) ? 0.5 : -0.5)) /    \
         scale_multiplier);                                                  \
diff --git a/cpp/src/gandiva/precompiled/extended_math_ops_test.cc b/cpp/src/gandiva/precompiled/extended_math_ops_test.cc
index fcbefe1bb306..fb51342dec45 100644
--- a/cpp/src/gandiva/precompiled/extended_math_ops_test.cc
+++ b/cpp/src/gandiva/precompiled/extended_math_ops_test.cc
@@ -93,6 +93,10 @@ TEST(TestExtendedMathOps, TestRoundDecimal) {
   EXPECT_FLOAT_EQ(round_float32_int32(-1234.4567f, 3), -1234.457f);
   EXPECT_FLOAT_EQ(round_float32_int32(-1234.4567f, -3), -1000);
   EXPECT_FLOAT_EQ(round_float32_int32(1234.4567f, 0), 1234);
+  EXPECT_FLOAT_EQ(round_float32_int32(1.5499999523162842f, 1), 1.5f);
+  EXPECT_FLOAT_EQ(round_float32_int32(static_cast<float>(1.55), 1), 1.5f);
+  EXPECT_FLOAT_EQ(round_float32_int32(static_cast<float>(9.134123), 2), 9.13f);
+  EXPECT_FLOAT_EQ(round_float32_int32(static_cast<float>(-1.923), 1), -1.9f);
 
   VerifyFuzzyEquals(round_float64_int32(1234.789, 2), 1234.79);
   VerifyFuzzyEquals(round_float64_int32(1234.12345, -3), 1000);