This research is focused on identifying urban sprawl pattern and extent in two rapidly growing major Ghanaian cities (Accra and Kumasi) and how urban expansion affected heat island effect over the period of 2002–2017 using remote sensing imagery. The research employed remotely sensed images from Landsat 7 and 8 missions for mapping the urban sprawl. Land cover classification was done by using object-based image analysis, and for land surface temperature estimation single-channel algorithm was used. The intensity and magnitude of urban heat island were estimated. The results showed that urban expansion was more dominant process than densification in both cities. A significant area of bare soils and sparsely vegetated lands became built-up accompanied by total disappearance of forest belt of Kumasi. The intensity and magnitude values indicated the presence and expansion of urban heat island in both cities. However, there was a significant amount of bare lands and sparsely vegetated areas with relatively high surface temperature in and around these cities. From the results of this work, we note that bare or sparsely vegetated land cover types in urban areas located in tropical climates can escalate overall urban temperatures. The urban heat island magnitude values were relatively higher compared to values for European cities during the heat wave of 2016. Although urban configurations and climatic conditions may be the reason for the differences, this shows how alarming and dangerous urban heat islands could be in tropical cities.