In the study of vehicle dynamics, the accurate description of tire mechanical characteristics is the basis and key of vehicle dynamics simulation. An innovative tire model is proposed based on fuzzy algorithm and a sinusoidal membership function is used to design fuzzy rules. In order to ensure the accuracy of tire behavior calculation, this model is driven by a small amount of experimental data of tire mechanical characteristics. This tire model consists of four layers of fuzzy systems, each of which has a knowledge base. The data in knowledge base I is obtained by experiments, and the data of knowledge base II is computed by the upper system, and so is the later system. Then, the input signal, the change rate of side slip angle and slip rate, is considered to improve the calculation accuracy of the model. The proposed fuzzy tire model can accurately predict the longitudinal force, lateral force and self-aligning torque of the tire under unknown conditions. Finally, by comparing the fuzzy tire model with the experimental data, it is found that the maximum RRMSE (Relative Root Mean Square Error) value is not more than 0.14. It is proved that the model can accurately describe the tire mechanical characteristics under combined conditions.