The intermodal transportation of goods by vessels has increased over the years. In this context, the Berth Allocation Problem (BAP) arises and becomes fundamental to guarantee the efficiency of the maritime terminals, deciding where and when to allocate the vessel over a planning horizon taking into account constraints of time and space. Because the problem is proved NP-hard, this study proposes an exact method and analyzes metaheuristics for tackling the problem. First, considering the BAP as a parallel-machine scheduling problem, an approach for this problem is proposed based on an Evolutionary Metaheuristic, aiming to Ąnd several good quality solutions in a single round of the algorithm, considering explicitly the BAP with multiple objectives. A lower bound based on a maximal flow problem was derived in order to evaluate the quality of the solutions. Next, based on a heterogeneous vehicle routing problem with time windows a basic Benders Decomposition algorithm and its variants are reviewed and applied to the BAP. Then, a hybrid optimization procedure based on Genetic Algorithm (GA) and Scatter Search (SS)is developed, and data envelopment analysis (DEA) is adopted to choose the efficient combination of the operators for the algorithm proposed. Because most papers in literature use in their experiments data generated randomly, making comparisons between researches difficult, this thesis proposes a problem generator for the BAP, allowing the generation of appropriate test problems to be commonly used with speciĄc desired properties and under controlled conditions. The data are generated using different parameters and the difficulty of solving the BAP with such data is analyzed through the resolution using the CPLEX. Finally, the instances classiĄed as more difficult are solved through two metaheuristics implemented.