Abstract—Vehicular ad-hoc networks is very popular research domain in which research work is going on at various aspects like routing the data without loss end-to-end. Routing in such networks is very tedious task due to frequently changing the position of vehicles location-wise. In this paper an intelligent model has been developed on the basis of adaptive neuro fuzzy system for OLSR routing protocol in VANET. The proposed model is designed based on input parameters average goodput and mac/phy-overhead. Based on these parameters, transmission power can be predicted. Triangular and Gaussian membership functions have been applied for designing the decision model. A comparison work also has been carried out for Gaussian, triangular functions and NS-3 based results. At the same time, the model is investigated by simulation work carried out on network simulator-3 (NS-3) platform.