A novel fuzzy sliding mode controller for robotic manipulators is proposed. The use of a boundary layer in sliding-mode control has been a common technique to reduce chattering of the control input. However, different choices of the boundary layer width lead to conflicting effects. We apply fuzzy logic to adjusting the width ofthe boundary layer based on the state. The method can effectively eliminate chattering and ensure almost peflect control accuracy at the same time. However, the intelligent control strategies have high control input requirement, which conjnes the practical application of modern control strategies in large extent So we add consrrained input control strategies in the controller; and present fizzy sliding mode constrained input controller based on systemic state. Simulations are given on a two-link robot in the end of paper; and the results show superior control performance than the conventional sliding mode controller for robot manipulators control.