diff --git a/docs/en/12-taos-sql/10-function.md b/docs/en/12-taos-sql/10-function.md
index 8ab5176aab0a3cd9a2dc4395febbf6ddc8f595f3..4504456cd7fb29a7147479065623c24d7516491f 100644
--- a/docs/en/12-taos-sql/10-function.md
+++ b/docs/en/12-taos-sql/10-function.md
@@ -795,19 +795,23 @@ HISTOGRAM(expr，bin_type, bin_description, normalized)
 ### PERCENTILE
 
 ```sql
-PERCENTILE(expr, p)
+PERCENTILE(expr, p [, p1] ...)
 ```
 
 **Description**: The value whose rank in a specific column matches the specified percentage. If such a value matching the specified percentage doesn't exist in the column, an interpolation value will be returned.
 
-**Return value type**: DOUBLE
+**Return value type**: This function takes 2 minumum and 11 maximum parameters, and it can simultaneously return 10 percentiles at most. If 2 parameters are given, a single percentile is returned and the value type is DOUBLE.
+                       If more than 2 parameters are given, the return value type is a VARCHAR string, the format of which is a JSON ARRAY containing all return values.
 
 **Applicable column types**: Numeric
 
 **Applicable table types**: table only
 
-**More explanations**: _p_ is in range [0,100], when _p_ is 0, the result is same as using function MIN; when _p_ is 100, the result is same as function MAX.
+**More explanations**:
 
+- _p_ is in range [0,100], when _p_ is 0, the result is same as using function MIN; when _p_ is 100, the result is same as function MAX.
+- When calculating multiple percentiles of a specific column, a single PERCENTILE function with multiple parameters is adviced, as this can largely reduce the query response time.
+  For example, using SELECT percentile(col, 90, 95, 99) FROM table will perform better than SELECT percentile(col, 90), percentile(col, 95), percentile(col, 99) from table.
 
 ## Selection Functions
 
diff --git a/docs/zh/12-taos-sql/10-function.md b/docs/zh/12-taos-sql/10-function.md
index c44c09d10c115a41efd22048d4fdafdc85370ac2..bc51a51209f59cc8941e147591427af20ae4ea44 100644
--- a/docs/zh/12-taos-sql/10-function.md
+++ b/docs/zh/12-taos-sql/10-function.md
@@ -798,18 +798,22 @@ HISTOGRAM(expr，bin_type, bin_description, normalized)
 ### PERCENTILE
 
 ```sql
-PERCENTILE(expr, p)
+PERCENTILE(expr, p [, p1] ... )
 ```
 
 **功能说明**：统计表中某列的值百分比分位数。
 
-**返回数据类型**： DOUBLE。
+**返回数据类型**： 该函数最小参数个数为 2 个，最大参数个数为 11 个。可以最多同时返回 10 个百分比分位数。当参数个数为 2 时， 返回一个分位数， 类型为DOUBLE，当参数个数大于 2 时，返回类型为VARCHAR, 格式为包含多个返回值的JSON数组。
 
 **应用字段**：数值类型。
 
 **适用于**：表。
 
-**使用说明**：*P*值取值范围 0≤*P*≤100，为 0 的时候等同于 MIN，为 100 的时候等同于 MAX。
+**使用说明**：
+
+- *P*值取值范围 0≤*P*≤100，为 0 的时候等同于 MIN，为 100 的时候等同于 MAX;
+- 同时计算针对同一列的多个分位数时，建议使用一个PERCENTILE函数和多个参数的方式，能很大程度上降低查询的响应时间。
+  比如，使用查询SELECT percentile(col, 90, 95, 99) FROM table, 性能会优于SELECT percentile(col, 90), percentile(col, 95), percentile(col, 99) from table。
 
 
 ## 选择函数