It is necessary to assess network video quality in real-time that depends on network transmission statement to subjective visual without any referenced video sequences. Focusing on distortion of videos caused by network packet loss, it is found that network packet loss in different frames with different decoding types causes different sustained effects of video distortion, and it exists significant association between visual feelings caused by network packet loss and movement activity of lost video content. Hence, a no-reference assessment model for video quality is proposed based on movement activity. This model analyzes the damage caused by the packet loss of the code stream received by client, and marks the damaged macro block to associate the damaged macro blocks′ distortion continuing effect and movement activity with videos' quality. Then, it completes the quality assessment of videos without reference of any original video sequence or video decoding. Experimental results show that this method has a small amount of calculation, a high real-time and a high degree of consistency with the subjective assessment results.