The purpose of this study was to find conclusions by using multiple stepwise regression, 46 tennis major male players at National Taiwan College of Physical Education were the subjects. The conclusions were drawn from the tests of the 21 given items of tennis fundamental skill and performance for 28 days. The multiple stepwise regression equation was generated from the test result in order to be used as a reference of enrollment standard effectively and objectively. After statistical analysis, there were conclusions as follows: 1.Every test reliability after the test result, all reached standard level respectively (P<0.01). 2.The fundamental skill test and sport skill performance (run match rankname)(total 21 items) for predicting the effect of tennis sport skill performance, after the 7 steps (the X09th - factor forehand smash point was taken out[sweep-out]), the involved factors were X20th factor double defeat elimination game; X19th skill action of combination; X12th service of distance; X10th fore- hand smash of distance; X05th fore-hand and back-hand ground-stroke score point. The factors caused the sig- nifican level of regression mode (F=806.454, P<0.01). The multiple correlation coefficient was significant, R=0.9874, and decisive coefficient was R2=0.9749, which means the 5 factors caused 97.49% of variance in criterion, and the increased measurement was 1.53%. Therefore, the result has a raw regression (forecast) equation as follows: Y=-7.6687+0.4380X �� +0.3798X ���� +0.0791X �� -0.0701X �� -0.5120X �N The standardized regression equation as follows: Zy=0.4357ZX �� +0.5575ZX ���� +0.1351ZX �� -0.0940ZX �� -0.0592ZX �N