The school bus problem routing (SBRP) is an important practical problem studied in combinatorial optimization of operational research. It is formulated through a set of stops, buses, schools and garage, where from these sets, we seek to create optimized routes to reduce the operating cost of the service. This paper presents a solution to the SBRP, using the GRASP applied to a real problem. This meta-heuristic is divided into two stages: the construction of a viable solution and followed by a Local Search procedure. These two phases are repeated in each iteration. In the construction phase, a greedy and random function is used to construct an initial solution. It is also presented compared to the widely used heuristic and good results in the literature.