the identification of hidden troubles of distribution transformers in low-voltage stations plays an important role in the safe and stable operation of power systems. Therefore, it is of great significance to detect the early hidden troubles of distribution transformers as soon as possible. In this study, the modified bat algorithm optimized support vector machine (SVM) was introduced into transformer trouble analysis and diagnosis. Due to the low solution accuracy and convergence speed of the traditional bat algorithm, long training time and easy to fall into local optimum, a modification method is proposed that introduces adaptive coefficients into the velocity equation, weighting factors into the position equation, and improved coefficients into the frequency equation method. The modified bat algorithm is used to optimize the two main parameters C and $\sigma$, which affect the classification accuracy of the support vector machine, to obtain the best parameter combination and establish a hidden trouble identification model. Then, the dissolved gas analysis (DGA) data is substituted into the model to obtain Transformer hidden trouble identification results. This scheme has faster convergence speed, higher solution accuracy and stronger robustness. Compared with the traditional bat algorithm, particle swarm optimization algorithm, genetic algorithm and simulated annealing algorithm to optimize the support vector machine (SVM) for obtaining hidden danger identification results, the final experimental results show that using the modified bat algorithm to optimize SVM has higher recognition accuracy, diagnostic accuracy can reach 94.45%, and good generalization ability.