The cell formation problem determines the decomposition of the manufacturing cells of a production system in which machines are assigned to these cells to process one or more part families so that each cell is operated independently and the intercellular flows are minimised or the number of parts flow processed within cells is maximised. In this paper, a tabu search heuristic—TSCF—that consists of dynamic tabu tenure with a long-term memory mechanism is presented to solve the cell formation problem. Test problems adopted from the literature and generated randomly are used to evaluate the performance of the proposed algorithm. In addition, two methods for quickly generating the initial solutions are proposed, namely the group-and-assign (GAA) method, and the random approach. Computational results indicate that the GAA method, accompanied by the TSCF algorithm can produce optimal solutions in less than or equal to 0.005 s for all small- and medium-sized problems. The proposed algorithm should thus be useful to both practitioners and researchers.