Abstract： The configuration of manufacturing cell was focused on taking into consideration alternative process routings and operation sequences of parts without predetermined number of cells, and a mathematical model was proposed with the objective of minimizing intercellular movements. An automatic clustering approach based on the improved discrete particle swarm optimization was proposed for the cell formation problem (ACPSOCF). The selfadaptive parameter for mutation was introduced to improve the diversity of particle swarm and determine the best number of cells automatically. The experimental results verify the effectiveness of the proposed approach on all test problems, which exceeds or matches the quality of the best solutions presented in the literature, without predetermination of the number of cells.