Aiming at the problem of low efficiency and large error in power transformer fault diagnosis, a fault diagnosis model of power transformer based on parameter optimization is proposed. Firstly, the fault feature of power transformer is extracted as input of least squares support vector machine (LS-SVM), and the fault type of power transformer is output. Then the LS-SVM is used to learn the fault diagnosis samples of power transformer, and the classifier of power transformer fault identification is constructed, and the chaotic particle is introduced. The parameters of LS-SVM are optimized by the group algorithm. Finally, the fault diagnosis of power transformer is simulated and tested. Test results show that this model can accurately identify various types of power transformer faults, obtain a higher accuracy of transformer fault diagnosis results, power transformer fault diagnosis speed, and the overall performance of power transformer fault diagnosis is better than other current power transformer fault diagnosis models.