The traditional detection methods of stress response of mutton sheep focus on the static blood test method, which cannot achieve real-time and continuous detection in the process of animal farming and transportation. A set of devices for dynamic detection of wearable biological and environmental signals in the process of mutton sheep transportation was designed. The heart rate and body temperature of mutton sheep during transportation were obtained by pulse wave sensor and infrared temperature sensor. The time domain analysis method was used to extract five typical characteristic parameters of transport stress characteristic parameters of mutton sheep, and their characteristics were analyzed in detail. The characteristic parameters MR, SDNN, RMSSD, pNN50 and CV tended to increase with the extension of transportation time, and were random due to changes in roads, which proved the changes in transportation stress. The average prediction accuracy of the stress state prediction model based on expert experience and supervised machine learning method was more than 89.81%. The results showed that through the development of wearable biosensing devices the pulse wave and body temperature can be monitored and the characteristic parameters of transport stress of mutton sheep can be extracted. The three states of transport stress can be effectively identified, the classification prediction management mechanism of transport stress of mutton sheep can be realized, and the problems such as low efficiency of transport supervision and quality control of mutton sheep can be solved.