The diagnosis of urban heat islands have demonstrated to be an important instrument to manage the urban space since the identification of intra-urban temperatures differences and of those in nearby rural areas may offer subsidies for measures intended to alleviate the magnitude of such heat islands. However, interpolations of air temperature records do not consider surface targets characteristics, and produce results that render difficult more localized interventions. The objective of this article is to present a set of procedures for the modeling of urban heat island, in order to estimate the air temperature, taking into consideration surface characteristics (land use, vegetation, and relief). For that, relations between airs temperatures measured by means of mobile transects, land use based on Landsat 8 satellite images, NDVI (Normalized Difference Vegetation Index) and hypsometry were processed. With that information, statistical relations were performed between the urban heat island intensity in locations where temperatures were recorded and the geographic parameters, such as vegetation percent, density of buildings and altitude. Results showed how the model generated estimated satisfactorily (r = 0.81) the air temperature in intra-urban and nearby rural spaces, making it possible to provide more localized interventions that can contribute to alleviate the atmospheric heat island, which, in the episode studied (06/30/2014) has presented a strong magnitude and 9ºC intensity.