diff --git a/datafusion/functions/src/math/round.rs b/datafusion/functions/src/math/round.rs
index 78016c0f52f71..f7c0cf1b6e1a6 100644
--- a/datafusion/functions/src/math/round.rs
+++ b/datafusion/functions/src/math/round.rs
@@ -630,15 +630,36 @@ fn round_columnar(
     }
 }
 
+trait RoundEven: num_traits::Float {
+    fn round_even(self) -> Self;
+}
+
+impl RoundEven for f32 {
+    fn round_even(self) -> Self {
+        self.round_ties_even()
+    }
+}
+
+impl RoundEven for f64 {
+    fn round_even(self) -> Self {
+        self.round_ties_even()
+    }
+}
+
 fn round_float<T>(value: T, decimal_places: i32) -> Result<T, ArrowError>
 where
-    T: num_traits::Float,
+    T: RoundEven,
 {
+    if decimal_places == 0 {
+        return Ok(value.round_even());
+    }
+
     let factor = T::from(10_f64.powi(decimal_places)).ok_or_else(|| {
         ArrowError::ComputeError(format!(
             "Invalid value for decimal places: {decimal_places}"
         ))
     })?;
+
     Ok((value * factor).round() / factor)
 }
 
@@ -809,6 +830,22 @@ mod test {
         assert_eq!(floats, &expected);
     }
 
+    #[test]
+    fn test_round_even_f32_one_input() {
+        let args: Vec<ArrayRef> = vec![
+            Arc::new(Float32Array::from(vec![2.5, 3.5, -2.5, -3.5])), // input
+        ];
+
+        let result = round_arrays(Arc::clone(&args[0]), None)
+            .expect("failed to initialize function round");
+        let floats =
+            as_float32_array(&result).expect("failed to initialize function round");
+
+        let expected = Float32Array::from(vec![2.0, 4.0, -2.0, -4.0]);
+
+        assert_eq!(floats, &expected);
+    }
+
     #[test]
     fn test_round_f64_one_input() {
         let args: Vec<ArrayRef> = vec![
@@ -825,6 +862,22 @@ mod test {
         assert_eq!(floats, &expected);
     }
 
+    #[test]
+    fn test_round_even_f64_one_input() {
+        let args: Vec<ArrayRef> = vec![
+            Arc::new(Float64Array::from(vec![2.5, 3.5, -2.5, -3.5])), // input
+        ];
+
+        let result = round_arrays(Arc::clone(&args[0]), None)
+            .expect("failed to initialize function round");
+        let floats =
+            as_float64_array(&result).expect("failed to initialize function round");
+
+        let expected = Float64Array::from(vec![2.0, 4.0, -2.0, -4.0]);
+
+        assert_eq!(floats, &expected);
+    }
+
     #[test]
     fn test_round_f32_cast_fail() {
         let args: Vec<ArrayRef> = vec![
diff --git a/datafusion/sqllogictest/test_files/scalar.slt b/datafusion/sqllogictest/test_files/scalar.slt
index 89ae30e3c047b..a4808cf5f95db 100644
--- a/datafusion/sqllogictest/test_files/scalar.slt
+++ b/datafusion/sqllogictest/test_files/scalar.slt
@@ -918,8 +918,8 @@ select round(a), round(b), round(c) from small_floats;
 ----
 -1 0 -1
 -1 NULL NULL
+0 0 0
 0 0 1
-1 0 0
 
 # round with too large
 #  max Int32 is 2147483647