In this work, we study the berth allocation problem (BAP), considering the cases continuous and dynamic for two quays; also, we assume that the arrival time of vessels is imprecise, meaning that vessels can be late or early up to a allowed threshold. Triangular fuzzy numbers represent the imprecision of the arrivals. We present two models for this problem: The first model is a fuzzy MILP (Mixed Integer Lineal Programming) and allows us to obtain berthing plans with different degrees of precision; the second one is a model of Fully Fuzzy Linear Programming (FFLP) and allows us to obtain a fuzzy berthing plan adaptable to possible incidences in the vessel arrivals. The models proposed have been implemented in CPLEX and evaluated in a benchmark developed to this end. For both models, with a timeout of 60 min, CPLEX find the optimum solution for instances up to 10 vessels; for instances between 10 and 65 vessels it finds a non-optimum solution and for bigger instants no solution is founded. Finally we suggest the steps to be taken to implement the model for the FFLP BAP in a maritime terminal of containers.