A method for calculating an error of a sampling clock is provided. The sampling clock is used for sampling a signal to generate a first sample data group and a second sample data group. Each of the first and second sample data groups includes a header having a predetermined sequence. The method includes: performing a correlation operation on the first and second sample data groups with data of the predetermined format to obtain first and second correlation results, respectively; comparing the first and second correlation results to generate a sample data group offset; and generating the error of the sampling clock according to the sample data group offset and a time difference between the first and second sample data groups.