For prediction of static and dynamic performance of doubly-salient motors, it is essential to know their flux linkage-position-excitation characteristics and also the static torque characteristics. At the design stage determination of these characteristics presents difficulties because of highly nonlinear behavior of the magnetic circuit. It is possible to use numerical field solution of the complete motor to obtain this information. This, however, requires expertise on a professional program and may be expensive if used to search for the best design. This paper shows that a reduced model can be used to obtain the desired information accurately. It is also shown that in fact obtaining field solutions just for a pair of teeth is enough for accurately predicting the flux linkage and torque characteristics of a motor. The approach introduced here makes possible searching for an optimum design (even on a PC) for maximizing average torque or reducing noise and vibration problems, since the effort for producing the model and computation time are greatly reduced.