The growing availability of data coming from ship reporting systems, such as Automatic Identification System (AIS) and Long Range Identification and Tracking (LRIT), is originating an unprecedented set of opportunities to enforce maritime surveillance, ensure the security of the traffic at sea, and manage maritime operations. In this paper, a data-driven methodology is proposed to estimate the vessel times of arrival in port areas. The developed approach exploits both AIS and LRIT historical maritime traffic data collected over a desired area of interest and is based on an optimized data-driven path-finding algorithm. The methodology is applied and validated to real scenarios with real data sets, showing how a list of times of arrival can be automatically computed for predefined ports and progressively refined. Such information is expected to increase port operational efficiency and safety.