The stands and soils of forest ecosystems are one of the most important carbon sinks in the terrestrial environment; thus, proper estimation of forest biomass increment is crucial to understand the rate of atmospheric CO2 sequestration. Here, we propose a biomass and biomass increment estimation methodology using forest inventory data from 1991 to 2006 from the Tatra National Park (TNP), Western Carpathians, Poland. Using machine learning techniques (random forest), we showed, for the first time, that stand age, the proportion of Abies alba, Picea abies, and Pinus mugo, and elevation are the most important predictors of aboveground stand biomass. Stand biomass increment is driven by the proportion of P. mugo, A. alba, and P. abies, stand age, elevation, and valley depth. Our results showed that the primary drivers of forest biomass (stand age and species composition) could be modified by the geomorphic properties of the terrain, indicating their importance in the mitigation of negative climate change trends in forest landscapes. Additionally, we showed that the biomass increment assessment using repeated measurements differed from trends of age-based biomass models, indicating the need to consider site-specific factors (e.g., slope and aspect). The potential applicability of these results improves the accuracy of carbon budget modeling and supports decision-making in nature conservation and forest management.