Traditional controllers normally need the mathematical model of the plant. But it is very hard toget the precise mathematical model of the plant or to get its result. 1n order to minimize the dependence on the plant information , Adaptivc Sliding Mode Control (ASMC) is applied to track and control a family nonlinear system. The scheme can function without any plant model. without expert knowledge , and without the previous plant data. which uses the radial based function neural networks to adjust the system input ,so as to ensure the stability and adaptation. Its global stability is proved through theoretical analysis. The simulation of the controller shows its strong robustness and feasibility.