This problem discussed in this work is similar to the combination of Vehicle Routing Problem (VRP) and Travelling Salesman Problem (TSP), however new contribution is proposed. In this work, a heuristic methodology is proposed to determine the set of the bus stop from the original set of the bus stops. A new approach is proposed to construct a matrix which is collecting the probability of the passengers willing to the next bus stop on the bus from all the bus stops. With this matrix, the next bus stop will be determined to travel to from the previous bus stop. The real distance between two corresponding bus stops is calculated by Dijkstra and Depth First Search algorithm and with some additional feature to propose a new bus route. This algorithm is named by MPBR (Most Profitable Bus Route by Flow Network). The bus transportation problem was implemented in San Francisco. The result is obtained by using this approach showed reducing the passengers waiting time