Adaptive resonance theory (ART1) network is one of the many popular neural networks used to solve the cell formation problem. Several modifications of ART1 for the problem have recently been published. In this study, a modified ART1 network is integrated with an effective optimization technique, Tabu Search (TS), to solve cell formation problems. The number of exceptional elements (EE) and group efficiency (GE) are considered as the objectives for the problems under the constraints of the number of cells and cell size. This proposed heuristic (ART1&TS) first constructs a cell formation using a modified ART1, and then refines the solution using a basic TS heuristic. ART1&TS has been applied to most popular examples in the literature. The computational results showed that it generated the best solutions in most of the examples.