This paper mainly researches on the identification issue through lower limbs. A new gait identification method is proposed by using the closed area of lower limbs as basic feature extraction. Firstly, skeleton is applied to obtain body skeleton, and through connecting lowest points of left and right foot, a closed area is formed as the gait feature. Then wavelet transform is used to analyze the feature datum. Finally, identification and classification of feature datum is realized by the mature BP neural network. The proposed method has been applied to the 0°camera angle of CASIA data-base, and the test results show the method's efficiency to improve the gait identification rate.