In this paper we present the concept of green routing or eco routing and how one would implement this using attributes defined in real time traffic information and digital road map. The routing referred to this paper applies mainly to road networks where one wants to get from point A to point B by minimizing the fuel consumed in a vehicle. The new model of speed profile prediction which yields stop and go speed at the start, end and zero of the acceleration under sub-node and link conditions is described. The results for the lowest fuel consumption show about 8.5% of simulated routes and 5% (9%) of vehicle test routes saving in fuel, when compared to that for the existing route planning course (recommended and short distance route), even traveling time about 10% longer time. These results clearly show prediction accuracy of fuel consumption between modeling and measurement is below the average about 12%.