In order to overcome the problems of low fault recognition rate, slow convergence speed and easy to fall into local extremum in the current traditional back propagation (BP) model for heat network leakage fault diagnosis, an optimized BP model based on genetic algorithm-ant colony optimization (GA-ACO) algorithm was proposed. The initial value of the pheromone is improved by mean of using the cross-variance operator of the GA algorithm, the iteration speed of the model and the search for the optimal solution are advanced by the ACO algorithm, the initial weights and thresholds of the BP model are optimized, and the model is applied to the diagnosis of heat network leakage faults by using the system simulation software. The results show that compared with the traditional BP model and GA-BP model, the GA-ACO-BP model possesses faster convergence speed and the predicted value is closer to the expected value with less error, which effectively improves the prediction accuracy of heat network leakage faults and enables fast and accurate diagnosis and localization of leakage faults.