Magnetic components are widely used in power electronic equipment. Accurate prediction of core loss is considered crucial for system efficiency optimization and thermal management. However, the traditional Steinmetz equation does not incorporate the temperature variable. It inherently fails to characterize the nonlinear relationship between core loss and temperature variations. Thus, inherent limitations are presented when the equation is applied under wide temperature range conditions. To improve the prediction accuracy of the Steinmetz equation, firstly, temperature was integrated into the model based on data analysis. The effect of temperature on core loss variation was estimated. Then, a simplex method was proposed for parameter optimization of the modified Steinmetz equation. After the temperature variable was designed and introduced, objective and error loss functions were constructed. Parameter optimization was performed using the simplex method. The reasonableness and robustness of the temperature-modified Steinmetz equation were verified through comparison between the error loss function results and the optimal parameters. Finally, sensitivity analysis confirmed that the introduction of temperature enhanced the significance and accuracy of the output from the Steinmetz equation.