In order to reduce the unbalanced load of large wind turbines produced by the blade of wind shear and tower shadow effects, according to the wind turbine gas dynamics, wind shear and tower shadow effects, this paper proposes an independent variable pitch control strategy based on RBF neural network sliding mode. The ability of sliding mode to control of anti-interference and robustness is strong and it has quick response speed, the disadvantage is that the control of sliding mode jitters easily. By the online learning ability of RBF neural network, it adjusts the gaining of sliding mode controller in real time, leading the synovial function on switching surface, reducing the jitter of the sliding mode control effectively, and improving the dynamic performance of independent variable propeller control system. A joint simulation mode joint for 5 MW wind turbine blade is set up by Matlab/Simulink and GH-blade software. Simulation results show that the independent variable pitch control scheme can effectively reduce the unbalanced load in blade root and improve the power of wind turbines run below the rated wind speed in performance. Through the test platform, it also verifies the rationality of the independent variable pitch control strategy proposed in this paper. This work is supported by International S & T Cooperation Program of China (No. 2011DFA62890).