Excessively high temperature during milling will shorten the life of the milling tools, reduce the surface quality of the workpiece, and increase production costs. In this paper, a novel cutting temperature prediction model for milling Al7050 is proposed, which considers both tool geometry parameters and milling parameters. The aim is to adjust the milling conditions in advance according to the predicted temperature, thereby prolonging the tool life and improving the machining quality. Through single-factor experiments, it is found that there is a polynomial relationship between tool geometric parameters and cutting temperature. Based on the empirical formula of cutting temperature, an improved mathematical model of cutting temperature is proposed. In this model, eight variables including tool geometry parameters and milling parameters are considered. The coefficients of the mathematical model are determined by the cyclic fitting method. The simulation data and experimental data show that compared with the temperature prediction model considering only milling parameters, this model has better prediction accuracy. This idea can also be used to establish the cutting temperature prediction model of other materials.