This paper deals with flexible flow line scheduling problem. Studied researches of the literature assumed that either no setup has to be performed or that setup times are sequence-independent, chiefly. Besides, they are mostly conducted in single objective environment. Therefore, in this study, these two real world concepts are mimicked in the flow line model. The considered objectives of the proposed model are minimizing makespan and total tardiness. Due to the hard solvability of the problem, this research proposes a new multi-objective algorithm based on the single objective tabu search, called gradual transition tabu search (GTTS). GTTS is implemented in two phases and in each phase it tries to gradually modify one objective to the other objective. The neighborhood search efficiency is also improved by defining new candidate list strategies and implementing new dynamic tabu tenure. In addition, the results of the proposed algorithm are compared with two popular existing algorithms, called fast non-dominated sort genetic algorithm (NSGA-II) and multi-objective genetic algorithm, on a number of randomly generated test problems. The suitability of GTTS is demonstrated through different multi-objective evaluation metrics.