To solve the optimal path problem of the traffic network, an improved travel time estimation model is proposed, and an optimal path algorithm based on this model is designed. The travel time estimation model is improved on the basis of the segment truncated quadratic velocity trajectory model by replacing the velocity measured at the same departure moment with the arrival velocity of the road segment nodes, and the travel time is estimated by constructing a velocity trajectory that is continuous in time and space. The optimal path algorithm based on travel time estimation firstly solves K shortest paths based on Yen"s KSP algorithm with road section distance as impedance, secondly estimates the travel time of K shortest paths by the improved travel time estimation model respectively, and finally selects the optimal path with travel time as cost. The validity and superiority of the model and algorithm are verified by numerical experiments of Sioux falls network. The experimental results show that: the improved segmented truncated quadratic speed trajectory model improves the accuracy by an average of 65% compared with the original model; the optimal path results based on the algorithm of this paper can reduce the number of intersections the path passes through and shorten the total length of the optimal path, Moreover, the estimated results of the optimal path"s travel time stay within 3% of the real value of MAPE. The results of this study provide a theoretical basis for the optimal path method for traffic networks.