The purpose of this study was to evaluate the smoothing error of data before and after impact in sports movement according to smoothing methods based on physical inspect. The used smoothing methods are cubic smoothing spline. digital filter and wavelet and the technique to separate data before and after impact and to smooth the each data separately and the technique to skip data in smoothing process. At first. Using the math functions. theoretical data were made. Each smoothing methods with various parameters were applied to the data which were made with the various speed ratio and the relative angle between before and after of impact. Next time we use the method with optimal parameter to the experimental data, according to the speed ratio and the relative angle between before and after of impact. The results of this study were as follows: 1. It is effective to separate data at impact point and to smooth each data. Especially this is expected to have good result for velocity before and after impact. But, padding technique needs the accuracy of digitizing data before and after impact. 2. Sampling speed of data was important to obtain the accurate results. More skipping digitizing data. the error is larger. But the result to use 5-data before and after impact and skip randomly in the other data was nearly identical to the result to use all data for smoothing. 3. Spline method and digital-filter method of variation of parameter were large according to the state of impact and the separation technique, skip technique. but it was small on wavelet method. So, wavelet method can be used to smooth various kinds of data without cautions.