The daily movement data of pigs, including exercise time, distance, speed, acceleration and so on, can be used as an important data base for the analysis and evaluation of the health status of pigs. Collecting movement data of pigs by sensors would make animals uncomfortable. Combined with video surveillance and digital image technology, a methodology for detecting and tracking pigs under natural conditions was put forward. Then state parameter and motion parameters of displacement, velocity, acceleration and angular velocity were extracted by shortest distance matching algorithm. Eight experiments in real-time monitoring of farms were given and analyzed. Experimental results show that the proposed method has good performance in real farm scene with adaption to mild adhesion and illumination change. In addition, the accumulated value of displacement, as well as velocity, acceleration and angular velocity of motion extracted from the videos, can reflect the movement of pigs in the overall trend and provide the data foundation for future research of pig behavior.