Abstract: Aiming at long-term overhead work, comprehensive ergonomic evaluation indicators are extracted and a comprehensive ergonomic evaluation model empowered by correlation analysis is constructed by using physiological indication data based on surface electromyography (SEMG) signal, joint pose indication data based on convolutional pose machine recognition algorithm and rapid upper limb assessment method (RULA) and subjective evaluation indication data based on Borg scale in this paper. Meanwhile, based on the fatigue accumulation principle of each element during operation, a comprehensive ergonomic prediction model based on convolutional neural network and Transformer model is constructed. Finally, taking the disassembly task of a certain type of vehicle-borne antenna array cover plate as the verification objects, the ergonomic evaluation software is developed by integrating the above methods to quickly calculate the comprehensive ergonomic score, which proves the usability of the evaluation model proposed in this paper.