This study developed a bottleneck-based heuristic (BBFFL) to solve a flexible flow line problem with a bottleneck stage, where unrelated parallel machines exist in all the stages, with the objective of minimizing the makespan. The essential idea of BBFFL is that scheduling jobs at the bottleneck stage may affect the performance of a heuristic for scheduling jobs in all the stages. Therefore, in BBFFL, a variant of Johnson's rule is used to develop a bottleneck-based initial sequence generator (BBISG). Then, a bottleneck-based multiple insertion procedure (BBMIP) is applied to the initial sequence to control the order by which jobs enter the bottleneck stage to be the same as that at the first stage. Five experimental factors were used to design 243 different production scenarios and 10 test problems were randomly generated in each scenario. These test problems were used to compare the performance of BBFFL with several well-known heuristics. Computational results show that the BBFFL significantly outperforms all the well-known heuristics.