It is widely known that mixing in reacting sprays is an important factor in the determination of the flame lift-off length, which is the location where the flame stabilizes. These details can be captured with computational techniques using computational fluid dynamics with a very fine mesh resolution. Use of a coarse resolution overpredicts mixing due to large numerical diffusion and thus reduces the predicted lift-off length. Even though a recently developed gas jet spray model improves the prediction of the spray structure with coarse mesh resolution, predictions of vapor mixing in evaporating sprays are still mesh-dependent. In this study, a subgrid-scale model based on classical jet theory is presented, where the vapor-air mixing is modeled with a combined Lagrangian and Eulerian approach. The vapor is transported as a Lagrangian particle consistent with jet mixing and transport theory until the jet mixing is resolved by the mesh scale. In this way, the results show improved predictions of vapor tip penetration and flame lift-off length using coarse-mesh resolutions. The new model offers a potential tool for investigating reacting sprays using coarser mesh resolution to save computational time for complete cycle simulations of internal combustion diesel engines.