This paper introduces a methodology to obtain optimal routes for the collection of urban waste. The problem is characterized as arcs coverage and for its development a combination of techniques in the Operational Research area was used. Firstly, the Teitz and Bart heuristic was used to obtain p-medians; from the definition of latter were defined groups (clusters) of demand points by designating these to the medians, by means of the Gillett and Johnson algorithm, adapted. Finally, from the definition of the groups of points, the Chinese Postman Algorithm was used, resulting in the routing (sequencing the demand points to be served) in each of the service groups. The results obtained are presented for a case study of the city of Joinville, State of Santa Catarina.