In this paper, the main objective is to develop a fast analytical model to be used in optimization process or coupled with other fast analytical models. A reluctance network of DC series machine is developed and shows its accuracy and robustness thanks to the consideration of the armature magnetic reaction (AMR) which is an important phenomenon in high saturated DC motors. This model is validated by a statistical way and presents a fast computation time thanks to an equivalent symmetrical machine. The optimization is used with RN and FE models and shows the efficiency of the RN to find an optimum without needing the FE final step optimization.