This paper presents an approach to estimate link travel time in urban areas. This approach consists of a data fusion from underground loop detectors and probe vehicles equipped with global positioning system (GPS). This method is expected to be more accurate, reliable and robust than using either of these data sources alone. In this approach, an algorithm is developed. This algorithm is based on the unscented Kalman filter using vehicle counts and flows from loop detectors located at the end of every link, and travel time from probe vehicles. From these counts the average travel time is calculated using the “cumulative plot” method. Furthermore, in order to incorporate the GPS data, a map-matching method is used to associate a travel time to the appropriate link.