Considering the influence of the mechanical transmission flexibility on the servo control system and its control parameters, and the feed drive dynamic performance, the ball screw feed system state space model was firstly constructed in order to integrate to servo control system. Then, this state space model was integrated with servo control system model based on the digital module simulation method. And the correctness of this integration model was verified by experiment test. The test result showed that the error between experiment curve and the simulation curve was small, and the simulation result was 50.45Hz,-1.78dB in the first order natural frequency while the experiment result was 47.12Hz, -4.46dB in the first order natural frequency. In view of uncertainty of servo control parameter optimal value resulting from the nonlinear of cascade control and its dynamic characteristic, feed drive dynamic performance optimization was limited to some extent by the method of servo control parameter optimization choice. Hence, the neural network adaptive current and speed feed forward control design strategy was put forward. Based on this ball screw feed drive integration modeling, servo control parameter timevarying optimal selection as well as current and speed feed forward control strategy were combined to realize the 〖JP3〗ball screw feed drive dynamic response optimization. The following error decreased from 0.2526 to 0.1115, and interference peak decreased from 0.019 to 0.0070, rising time decreased from 0.1283s to 0.1075s. It can be seen that the response speed of ball screw feed drive system was faster while the following error was smaller, and the system antiinterference ability was better at the same time. The research of this paper provided the important theoretical basis and method guidance for the study of the overall dynamic performance and optimization of the ball screw feed drive system.