This research develops a new methodology for a cell formation problem in the cellular manufacturing system. The methodology includes two phases. In the first phase, an improved similarity coefficient method, which considers the operation sequence and the number of repeated operations firstly in related researches, is proposed to identify part families. A new decomposed mathematical model is presented in the second phase, which considers some crucial operational aspects such as alternative routing, machine capacity, part demand, operation time, and lot splitting, to assign machines into part families for minimum machine cost, operation cost, and inter-cell movement cost. The model puts emphasis on the effect of trade-off between machine duplication and material inter-cell movement on performance of the cell formation to optimize machine utilization and workload balance. This paper also provides a concrete production schedule with optimum system utilization for cell formation. Test problems and sensitivity analyses are carried out to reveal the effectiveness and feasibility of the proposed methodology.