In order to understand the complex nonlinear relationship between the influencing factors and the forces, the vertical sliding (braking and driving), side bias and complicated working conditions were carried out on the KHAT low-speed flat tire test-bed respectively. Furthermore, under the complicated working conditions, the influence of the factors such as tire pressure, vetical load, tire speed, sideslip angle and side rake angle on the lateral force as well as the longitudinal force are analyzed qualitatively. In the light of complex nonlinear relationship, a BP neural network tire mechanics model is established. The model is trained by experimental datas and compared with the "Magic Formula" tire model. The results show that the established BP neural network model can better approximate the test curve, and verify the validity of the model. It provides a reference for the study of tire mechanical properties.