diff --git a/be/src/exprs/function/function_split_by_regexp.cpp b/be/src/exprs/function/function_split_by_regexp.cpp
index d889b0fc8622b8..292bcb64559747 100644
--- a/be/src/exprs/function/function_split_by_regexp.cpp
+++ b/be/src/exprs/function/function_split_by_regexp.cpp
@@ -27,6 +27,7 @@
 #include "core/types.h"
 #include "exprs/function/function.h"
 #include "exprs/function/simple_function_factory.h"
+#include "util/simd/vstring_function.h"
 
 namespace doris {
 
@@ -115,7 +116,8 @@ bool RegexpSplit::get(const char*& token_begin, const char*& token_end) {
             }
         }
 
-        _pos += 1;
+        const auto utf8_byte_length = get_utf8_byte_length(static_cast<uint8_t>(*_pos));
+        _pos += std::min<size_t>(utf8_byte_length, _end - _pos);
         token_end = _pos;
         ++_splits;
     } else {
diff --git a/regression-test/data/query_p0/sql_functions/string_functions/test_split_by_regexp.out b/regression-test/data/query_p0/sql_functions/string_functions/test_split_by_regexp.out
index fa6e31abd94054..d9184915ab7b0e 100644
--- a/regression-test/data/query_p0/sql_functions/string_functions/test_split_by_regexp.out
+++ b/regression-test/data/query_p0/sql_functions/string_functions/test_split_by_regexp.out
@@ -2,6 +2,12 @@
 -- !select1 --
 ["a", "b", "c", "d", "e"]
 
+-- !select_utf8_empty_pattern --
+["中", "a", "😀"]
+
+-- !select_utf8_empty_pattern_with_limit --
+["中", "a😀"]
+
 -- !select2 --
 ["a", "bc", "de", "f"]
 
diff --git a/regression-test/suites/query_p0/sql_functions/string_functions/test_split_by_regexp.groovy b/regression-test/suites/query_p0/sql_functions/string_functions/test_split_by_regexp.groovy
index 678f72216421ce..9754ca6e592f3a 100644
--- a/regression-test/suites/query_p0/sql_functions/string_functions/test_split_by_regexp.groovy
+++ b/regression-test/suites/query_p0/sql_functions/string_functions/test_split_by_regexp.groovy
@@ -17,6 +17,8 @@
 
 suite("test_split_by_regexp") {
     qt_select1 "select split_by_regexp('abcde','');"
+    order_qt_select_utf8_empty_pattern "select split_by_regexp('中a😀', '');"
+    order_qt_select_utf8_empty_pattern_with_limit "select split_by_regexp('中a😀', '', 2);"
     qt_select2 "select split_by_regexp('a12bc23de345f','\\\\d+');"
     qt_select3 "select split_by_regexp('a12bc23de345f',NULL);"
     qt_select4 "select split_by_regexp(NULL, 'a12bc23de345f');"