In 1990, The Japanese National Police Agency (NPA) planned a nationwide two-year research project including field experiments, “Travel Time Estimation and Prediction” aimed at realizing a reliable method of providing travel time information to drivers (1)*. The project was conducted by the Japan Traffic Management Technology Association and three system manufacturers (Sumitomo Electric, OMRON and Matsushita Communication) collaborated on research and experimental work. Several methods of estimating and predicting travel time were developed and evaluated in three major cities (Tokyo, Yokohama and Osaka). Four models, Sandglass, Delay Time, Auto Regressive and Neural Network Models were developed and evaluated using detector and license plate reader information, and field survey data in the three major cities. The first year evaluation results are described here.