This paper combines the regression model based on nonlinear least partial squares and the decision tree model to more comprehensively excavate the influencing factors of the shooting percentage of basketball games from the perspective of both basic data and higher-order data. It also calculates the weight coefficients and comprehensive scores of the indicators of the four teams’ shooting techniques in the restricted area through the fuzzy comprehensive evaluation method and studies the hitting rate of different shooting methods. After analysis, the regression coefficients of core strength training and core strength training time are 1.325 and 1.086 in the basic data and high-order data, respectively, which have a significant positive influence on the winning and losing of the tournament, and the decision tree intuitively demonstrates that the core strength training is a key factor in the hitting rate of the game shooting. The weight coefficient of “rebound” is the highest, with a weight coefficient of 20.0271, which has the greatest influence on the shooting percentage, and “traveling backhand shot” has the least influence on the shooting percentage.