In order to realize the accurate recognition of the license plate characters under conditions of complex illumination and the character with rotation, occlusion or fouling, a new License plate character recognition algorithm based on Improved Hidden Markov features is proposed, in which the recognition efficiency is improved by reducing the dimension of the features using the fast independent component analysis and By choosing representative training samples to participate in classifier training through which Reduces the requirements for hardware performance. The experimental results show that，the proposed algorithm can significantly reduce the running time and improve the recognition accuracy under the condition of keeping the original classification and recognition performance.