This paper explores the temporal composition of the main Land-use/Land-cover (LULC) trajectories, examines the spatial configuration of the trajectories, and derives the probabilities of transitions in the Northern Ecuadorian Amazon (NEA). This research uses a time-series of classified Landsat images that ranges from 1974 to 2002, and a set of spatial socioeconomic, demographic, and accessibility data assembled in a geographic information system. The LULC trajectories are analyzed for the Northern Intensive Study Area (NISA) using image algebra, and for the whole region, the NEA, using cluster analysis, landscape ecology principles, and spatial logistic regression models. In general, the trajectories are dominated (i.e., in terms of area) by recent transitions that contain forested classes (i.e., primary forest or succession), as well as the consistent representation of pasture through time. This exploratory analysis of LULC transitions suggests a set of clusters that form a "core and periphery" pattern in the NEA. This research shows how these clusters and probabilities of change can be used to characterize trajectories of LULC in the region.