However, conventional BP neural networks have many shortcomings, such as slow convergence, sensitivity to initial values and easily getting into local minimum. Meanwhile, the color correction model generated by the BP neural network method is not unique. Asa result, this paper brings about improvements to traditional neural network by adopting the mind evolutionary computation (MEC)and adaboost algorithm (Adaboost) to optimize the BP neural net-work, and hence proposes the MEC-BP-Adaboost neural network.The proposed algorithm can effectively avoid BP neural network getting into local minimum and achieve relevant results after each training