Aiming to realize the automatic identification of individual behaviors of breeders in this natural mating cage, reduce the systematic error of observation and sampling, analyze breeding chicken behavior and understand the health condition of chicken group, a realtime monitoring system based on nineaxis acceleration sensor and bluetooth wireless transmission was designed to automatically monitor and quantify the individual behavior of breeding chickens in this cage. Wavelet sym was used to reduce noise influence on original data preprocessing. According to the fluctuation of different behaviors acceleration curve, the characteristic value of acceleration data was extracted and the Kmeans clustering algorithm was used to identify behavioral features to obtain a stable clustering center. Distance comparison of acceleration sensor and performance test of portable battery and lithium battery as power supply equipment was done. Meanwhile, video was used to monitor and verify the five behaviors of breeding chickens. The system can quickly and continuously collect individual behavior information of breeding chickens in this cage, accurately identify individual feeding, drinking, fighting, mating and wingflapping behaviors. The average recognition accuracy of feeding was 94.31%, the average recognition accuracy of drinking was 92.53%, the average recognition accuracy of fighting was 84.03%, the average recognition accuracy of mating was 72.00%, and the average recognition accuracy of flapping was 92.31%. It was helpful to acquire individual behavior of breeding chickens without damage and quickly, and provide a scientific basis for optimal design and efficient management of facilities and equipment.