In past centuries, the surrounding countryside of fast-growing Andean cities have experienced a dramatic decline in forested areas. However, there is also evidence of forest recovery in some areas in recent times. Changes in land uses and cover have influenced not only forest cover extent but also forest fragments’ composition and ecosystem functioning. The main aim of this work is to gain a deeper understanding of the effects of anthropogenic perturbations on forest composition and cover dynamics in the Eastern Andean Cordillera of Colombia. Here, we pair an analysis of high-resolution aerial imagery covering the period 1940 to 2007 and plot-based community data obtained from six administrative localities. To elucidate land cover changes quantitatively through time, we performed a series of supervised land cover classifications and connectivity analyses to investigate the quality of forest surface changes. In parallel, plant communities were characterized, gathering data on both the woody vegetation and the understory compositional and structural parameters, while compiling a broad array of variables related to anthropogenic disturbance, ranging from local to landscape-wide metrics. A variable selection process was applied to identify the most influential predictors of community composition and structure by means of nonmetric multidimensional scaling (NMDS) and partial redundancy analysis (pRDA), and relevant predictors were used to build General Linear Models (GLMs) to assess their link to diversity metrics and aboveground biomass (AGB) values. The observed forest cover trend and the local recent history are discussed in relation to the compositional and structural parameters examined in each locality. This study furthers our understanding of High Andean Forest composition and extent in response to anthropic disturbance. Our results could be useful in guiding conservation measures to protect these delicate ecosystems.