The morphological configuration of cities has a direct influence on microclimatic variation. The predominant built composition in certain areas, the high rates of waterproofed surfaces, vegetation scarcity and water surfaces can have a significant impact on the temperature and humidity values on inhabited areas, often exposing the population on unhealthy environments. The aim of this paper is to elaborate a prediction microclimatic model variation of four different Local Climate Zones (LCZ) exposed to a dry atmospheric condition in an altitude tropical region. The method was developed on three stages, the first one, refers to a collect campaign of temperature and humidity variation in four different urban environments, LCZ D, LCZ 1, LCZ 5 and LCZ 9 in São José do Rio Preto, Brazil. The representative microclimatic behavior of each area of the city in relation to the performance of a dry air mass was recorded. The second stage involved the microclimatic collected data, which were submitted to a statistical analysis with ANOVA tests, serving as the basis for the development of prediction microclimatic models variation for each LCZs. After validating the models, it was verified, in the third stage, the urban area that presented morphological characteristics that allow the occurrence of high temperature waves and reduced indexes of relative humidity. The produced models for predicting urban microclimatic show a very high capacity of representation to estimate the temperature values in different areas of the city, since in all cases values of R exceed 80%. The results showed that the area with LCZ5 presents the longest periods of heat exposure, which should receive more attention from planners in relation to investments in urban green infrastructure.