Radiative transfer models are used in the routine analysis of huge amounts of Visible/Infrared satellite data. One of the application areas is the study of canopy reflectance (CR) to retrieve the amount, structure and functioning of vegetation canopies. Present CR models take into account the hot spot effect only in the first-order scattering of direct solar radiation, resulting in a violation of energy conservation. Recently, Kallel et al. [4,5] have proposed a model coupling the models Adding and SAIL for Discrete homogeneous medium, the AddingSD model. AddingSD takes into account the hot spot effect for diffuse fluxes and therefore ensures energy conservation. In our study, we propose the enhancement of this model as follows: (i) the enhancement of the first order hot spot effect Kuusk' model, namely the dependency on the leaf size, that the solution is injected in our model; (ii) extending AddingSD to heterogeneous media thus allowing to model forest canopy reflectance; and finally (iii) validating the new model using remotely sensed data (satellite and airborne) over Järvselja forest test site in Estonia. For airborne data, atmospheric correction algorithm will be developed to eliminate the effect of cirrus clouds present at measurement time.