Majority of currently applied environmental models relay on linear relations between environmental endmembers. In this research, a detailed and comprehensive comparison between linear (L) and non-linear (NL) models are presented. The L and NL models are realized in a framework of Artificial Neural Networks (ANN). The evolution process of the ANN-L and ANN-NL models is based on estimation of fractional snow cover through data-fusion between high resolution (IKONOS) and medium resolution (Landsat TM/ETM+) remotely sensed images. The statistical measure values of R2, RMSE, MAE, Accuracy, Precision, Recall, and Specificity indicate better performance of the ANN-NL model in comparison to the ANN-L in estimation of ANN Landsat-FSC. The presented results, strongly indicate that to fully capture, untangle, and characterize internal environmental relations high-sensitivity non-linear models are required. Non-linear relations are particularly visible the complex in alpine-forested environments.