In order to identify human behavior classification in Intelligent Security Monitoring System, an articulated model to extracting human body and classifying the behaviors of the moving objects is presented in this paper. An improved statistical Gaussian model is used as adaptive background updating method. After silhouettes of objects are extracted from the video images, we propose an articulated model of human, using the variety of body's trunk and limbs contour angles. The angles that can represent the pose of the skeleton model and length-width ratio of the human are used as feature vector. Finally Bayesian Networks is used for human posture training, modeling and activity matching to recognize the human motion. Experiment results have shown that this method gives stable performances and good robustness.