To solve the control problem of complex nonlinear system, a sliding mode predictive control based on lazy learning(LL-SMPC) is proposed. The LL-SMPC builds the local model on-line based on lazy learning algorithm and obtains the optimal control law by solving the quadratic optimization problem formulated in sliding mode predictive control framework, therefore it has strong adaptive and anti-jamming ability. Furthermore, the searching efficiency is improved by using hierarchical searching strategy, and the computation complexity is decreased by avoiding the solving of Diophantine equation. Simulation results show the effectiveness of the proposed method.