This study aimed to develop a multiple linear regression (MLR) statistical model to analyze the phenomenon of Urban Heat Islands (UHI) in the city of São Paulo, Brazil, between 2019 and 2023, using a mixed-methods approach. The research followed the Design Science Research (DSR) methodology, guiding all stages of the study, from defining the research framework to constructing the model. Additionally, both national and international literature, including studies from Málaga, Spain, and Stockholm, Sweden, provided the theoretical foundation for this dissertation. Meteorological, environmental, and health data were collected from institutions such as INMET, IAG-USP, CETESB, and the São Paulo Municipal Health Department. The analysis considered three strategic áreas (Area 1 – A1, Area 2 – A2 and Area 3 – A3), evaluating variables such as air temperature, relative humidity, precipitation, wind speed, and pollutant emissions. The study revealed that the more urbanized regions (A1 and A2) exhibited higher thermal increases and health impacts, such as a rise in respiratory diseases. For instance, in January 2019, A1 recorded the highest average maximum temperature (32.41°C), while A3 registered the lowest in June 2023 (15.16°C), highlighting the role of vegetation in mitigating heat islands. The use of geoprocessing with satellite imagery enabled the creation of comparative maps, demonstrating the increase in land surface temperature between 2019 and 2023. Furthermore, the statistical model indicated that relative humidity, pollutant emissions, and precipitation were the most significant factors influencing temperature variation. This research reinforces the need for sustainable urban planning strategies, incorporating vegetation, shading, and natural ventilation to mitigate the UHI effect. Moreover, the developed statistical model has the potential for generalization and application in other urban contexts, contributing to public policies aimed at climate resilience.