Urbanization is a key issue when studying scenarios for sustainable development for the Amazonian region. This paper presents the potential of high-resolution imagery, RapidEye and CBERS 4, and fieldwork data to analyze the urban sprawl of three cities - Itaituba, Mojuí dos Campos and Alter do Chão - in southwestern Pará from 2011 to 2015. PAN CBERS4 multispectral and panchromatic images were combined using Principal Component Spectral Sharpening method. RapidEye images were used to map the urban occupation in 2011 by visual interpretation. Urban sprawl was identified from the interpretation of the CBERS 4 images of 2015. Ancillary data and fieldwork information complemented and characterized theses areas of the horizontal expansion of cities associated the urbanized allotments, housing estates, condominiums and informal spontaneous occupation. The results show that the horizontal expansion of these three cities has a tendency for area fragmentation within the urban occupation areas, as observed by the empty spaces interspersed among expansion areas. Economic factors, expectation of the implementation of large enterprises, the price of land, and the definition of the legal tenure of urban land may be linked to the recent expansion of these cities. The analysis of the urban space production, based on multi-temporal remote sensing high resolution data, provides information that can support a broader discussion about development models and urbanization planning for the Amazon.