Human action recognition is a hot research topic in the field of computer vision

especially in intelligent household. Due to action feature extraction is affected by various aspects of factors like environment and diversity of action itself

it makes action recognition becomes more difficult

so KINECT camera is used for feature extraction

and action descripting and optimizing of the extracted feature data are carried out. Neural network is used to train feature data to make this method achieve good performance. Comparative experimental results verify the effectiveness of this method.