This paper proposed a new face detection method which based on the Adaboost and the face skin color detection. Firstly，face images are mapped from ＲGB color space to the YCbCr color space，and thenthe skin color model is established to obtain the similarity of the face，the face region of the candidate isobtained by the morphological processing. In the training stage，the expansion of the target weight is restrictedby adjusting the weight of the error distribution. The phenomenon，training degradation and over adapt the distribution of the training target weight，is suppressed by modifying the renewal of the weight. The improved Adaboost algorithm is used to detect the candidate region of the face，which can improve the detection speed. The experimental results show that the proposed algorithm can effectively improve the detection rate and the detection speed，which can restrain the over adaptation of the training object.