In order to solve the problem of distortion of thermal parameters of mass concrete in bridge engineering, an inverse analysis method of thermal parameters of mass concrete based on uniform design theory and BP neural network was proposed. This method uses BP neural network to establish the non-linear relationship between the temperature field and thermal parameters of large-volume concrete; training samples of BP neural network are determined by uniform design method; during the training phase of BP neural network, additional momentum method is used to optimize the network structure; The error curve before and after optimization and the analysis results of multiple training processes show that the additional momentum method can significantly shorten the network training time, and the average absolute percentage error value and root mean square error value of the multiple training processes are stable. The adiabatic temperature rise, reaction rate constant, and thermal conductivity were inverted during the construction of the bulk concrete of the supporting platform of the saddle pier of the Taihong Yangtze River Bridge. The calculated temperature based on the inversion value agrees well with the actual measured value, and the maximum temperature peak error is only 1.1 ℃. Therefore, the inverse analysis method for thermal parameters of large-volume concrete based on uniform design theory and BP neural network is feasible and the inversion process is stable and convergent, and the inversion accuracy is high. It can be used to guide the temperature-controlled construction to reduce the risk of large-scale concrete cracking.