This paper presents a new solution approach for optimal substation expansion planning (SEP) within electric power distribution networks. A modified fuzzy membership matrix as well as a memorable cost index vector is introduced to find the optimal substation service areas. Besides these, a Learning Automata-based algorithm is introduced for simultaneous determination of optimal service areas and capacities of the distribution substations. Electrical constraints such as voltage drops, power flow, radial flow constraints, as well as all prevalent cost indices are taken into consideration. The developed method is conducted to solve the distribution substation allocation problem for an actual distribution network with about 200 000 customers, and obtained results are compared to those of other methods. Detailed numerical results and comparisons presented in the paper show that the proposed solution approach could noticeably improve the quality of problem solutions with low computational burden and can be used as an effective tool for SEP in large distribution networks.