Flexible job shop scheduling problem (FJSP) is one of the complex and important problems in operations management. Over time, many heuristic and metaheuristic approaches have been used to find solutions of FJSP. However, the potential of the gravitational search algorithm (GSA) is still unknown while solving FJSP. Therefore, the present work uses real number encoding-based GSA (RGSA) and chaotic GSA (RCGSA) to solve FJSP. The results of 35 benchmark problems and one industrial test case show that RCGSA performs significantly better than RGSA regarding the quality of solutions and convergence in a limited number of iterations.