This article presents a comprehensive method to predict the electromagnetic losses in surface-mounted permanent-magnet (PM) motors based on a hybrid analytical model (HAM). HAM will significantly improve the calculation speed compared with finite-element method (FEM) while keeping great accuracy, making it a competitive alternative for the analysis and optimization of PM motors. As the accurate field distribution is the bias of electromagnetic losses prediction, the magnet loss, sleeve loss, and copper loss can all be accurately obtained from their vector potential distribution using HAM with small computational burden. As for iron loss, the improved Jiles–Atherton (JA) model is proposed to build the relationship between the loss and the magnetic field in the iron region from the perspective of energy conversion. The predictions of magnet loss, sleeve loss, and copper loss using HAM agree well with the FEM. The experiment on surface-mounted PM prototype demonstrates the high accuracy of iron loss calculation using the combination of the improved JA model and HAM.