Routing problems involving heterogeneous vehicles naturally arise in several civil and military applications due to fuel and motion constraints of the vehicles. These vehicles can differ either in their motion constraints or sensing capabilities. Approximation algorithms are useful for solving these routing problems because they produce solutions that can be efficiently computed and are relatively less sensitive to the noise in the data. In this paper, we present the first approximation algorithm for a 2-Depot, heterogeneous vehicle routing problem when the cost of direct travel between any pair of locations is no costlier than the cost of travel between the same locations and going through any intermediate location.