In this paper two SAR images (acquired by different frequency) are integrated using a methodology called "multisource classification". The technique allows the generation of different classification scenarios (classification and reliability map). This different scenarios were combined in order to obtain better classification results by using the minimum function compounding the scenario multisource minimum. One RADARSAT-2 (C-band) and one ALOS/PALSAR (L-band) images were used in this study. These images were modeled by the complex Wishart or multi-look intensity pair distributions. From the fourteen generated multisource scenarios, ten showed improvement greater than 10% related to the corresponding individual ratings. It was noted that for these scenarios both images had been modeled by the same distribution and for four remaining cases each data were modeled by a specific distribution. Two multisource scenarios did not presented overall accuracy and kappa coefficient higher than the individual classification however they exhibited high values of accuracy for Intermediate Regeneration class. The results showed that our method is effective to improve the classification accuracy indexes when SAR images are multisource integrated.