The computation method of Cu-laser model is analyzed and described and a way that eliminates the unreasonabe results in a normal program is suggested.