Abstract: Virtual manufacturing cells is designed to modify logistics routes and optimize manufacturing cells, enhancing workshop production efficiency and logistical flexibility. They are widely used in multi-variety and small-batch production workshops. With new order insertion, changes in product types and process paths can lead to imbalances in equipment load and decreased production efficiency within original virtual manufacturing cells. To address this issue, a new indicator to measure the similarity between new orders and original virtual manufacturing cells is introduced, incorporating continuous identical operations. A mathematical model for virtual manufacturing cell reconfiguration is established with the objectives of equipment load balancing, total transportation cost, number of inter-cell movements, and cell inheritance rate. In addition, an improved multi-objective evolutionary algorithm based on decomposition (MOEA/D) is proposed, which utilizes multiple crossover operators to improve global search efficiency and adopts a simulated annealing-based neighborhood search to enhance local exploration capability. Finally, the effectiveness and excellence of the proposed method are verified by using an actual case from the machining workshop in a factory and 10 benchmark cases, providing theoretical references for decisions of virtual manufacturing cell reconfiguration.