The most common models for estimating rainfall interception are physical-based models, of which Gash models are the most widely used. Therefore, this study aimed to compare the efficiency of revised Gash models in estimating the rainfall interception of Fagus orientalis Lipsky in Siahkal forests, Gilan province. During the one-year measurement period in the study plot (March 2015 to March 2016), gross rainfall, throughfall, and stemflow values were collected and the amount of rainfall interception was computed. For modeling rainfall interception, Gash-1 and Gash-2 models were selected, and the performance of these two models was evaluated during time scales: annual, leafed, and leafless periods. The results showed that in general, canopy and trunk echohydrological parameters were obviously different amongst time scales. In all time scales studied, the Gash-2 model performs better than the Gash-1 model, and the most appropriate performance of Gash-2 model is in estimating annual interception (estimation error: +8.8%; efficiency coefficient: 0.92). In summary, it can be said that forest ecohydrological processes in deciduous broadleaf stands should be studied separately during different time scales, because the ecohydrological response of canopy cover in different time scales is varied. Also, different models in different time scales do not have the same performance. Hence, further studies are needed to determine the efficient models in each time scale of each forest stands, and by obtaining the appropriate model, we can help to accurately estimate the interception in Hyrcanian forests.