Amazonian rivers are highly interconnected and dynamic systems. Their behavior depends, to a large extent, on their geomorphology, being classified in 1) meandering rivers (MR), characterized by high rates of migration and sinuosity, and 2) anabranching rivers (AR), known for forming several permanent channels and islands. A planimetric characterization of the main rivers of the Peruvian Amazon (Huallaga, Ucayali, Marañon, and Amazonas), spanning from the Andes to the Amazon lowland region, was carried out to understand their physical dynamics. By a multi-temporal analysis from 1987 to 2017 using Landsat images, a segmentation was made for each river based on 1) the characterization of the geological valley, 2) the confluence of important tributaries, 3) changes of the main channel through the years, and 4) planimetric variables such as confinement, bend length, amplitude, sinuosity, and asymmetry. As a result, a total of 160 sections were obtained, in which a new set of 25 metrics was applied, filtered from an initial set of 31 variables and their statistics (i.e. mean, variance, kurtosis, and skewness), calculated through different approaches (i.e. half-meander, full-meander, and full-river). The variables were standardized and principal component analysis (PCA) was performed. The resulting biplot showed a distinction between AR and MR, with a shared area consisting predominantly of Marañon and Huallaga sections. The average value of sinuosity was found more associated with the MR, while higher length and asymmetry variance values were more oriented to the AR. This study also indicated the similarity in the behavior of some river sections of different types, based exclusively on their morphometric characteristics. At the same time, revealed how some sections could not be differentiated from others despite being nominally different. In this scenario, the PCA highlighted the need for a complete set of statistics that can recognize different features of these rivers, capturing greater complexity. Thus, the evaluation and segmentation of these planimetric variables, according to their planform characteristics, allows a better understanding of their dynamics, providing accurate information for coherent decision-making.