A new method of fault section location and isolation in distribution networks based on immune algorithm (IA) is presented. Immune algorithm gets individual evaluation from its affinity. Selection of the crossover and mutation individuals is not only based on its affinity but also concentration, which increases colony diversity, avoids partial optimization, and speeds up the rate of evolution. Besides, this method can establish memory units and retain part of the optimal solutions in order to avoid degeneration in the process of crossover and mutation and guarantee the result is the individual of the highest level of affinity when the algorithm is terminated. A method based on the idea of area division and fault alarm of the incoming break is proposed to improve the speed of location for open-loop operational distribution networks. Finally the simulation results show that the immune algorithm can realize the goal of fault location and the comparison results with genetic algorithm testify its validity and reliability.