This paper introduces a new indirect rotor position estimation algorithm for the SRM sensorless control, based on the magnetizing curves of aligned and unaligned rotor positions. Through the basic test method, the complete SRM magnetizing characterization is first constructed using a neural network training, and then used to estimate the rotor position. And also, the optimal phase is selected by the phase selector. In order to verify this approach, the proposed rotor position estimation algorithm using a neural network learning data is investigated. The experimental results show that the proposed control algorithm can be effectively applied to SRM sensorless control.