The Urban Heat Island (UHI) phenomenon results in higher temperatures in urban areas compared to less urbanized regions. This is due to the concentration of urban infrastructure, which absorbs and then releases solar radiation. Given its significant role in exacerbating the climate crisis, the UHI phenomenon demands urgent attention. While traditional physics-based simulations for studying UHI are accurate, they require substantial resources, which limits their practical application in urban planning. Previous research by the authors highlighted the capability of data-driven models as a practical alternative for assessing UHI. Such models, however, depend on the availability of extensive high-resolution datasets. Building on this prior work, the current study explores utilizing UHI’s seasonality to narrow the required data scope for effective data-driven UHI modelling. By strategically targeting data collection on specific seasons, it is possible to capture UHI’s intricate and dynamic nature more efficiently. This approach involved using street-based clustering to identify common seasonal patterns in Surface UHI (SUHI) and Canopy UHI (CUHI). Findings show notable seasonal fluctuations in SUHI, especially during summer. The training of Random Forest (RF) models employed varying data set proportions: 45% for summer and spring, 65% for autumn, and 75% for winter. Despite the challenges of smaller training datasets, the models achieved high accuracies, with CUHI models attaining an R 2 of 0.85 and SUHI models an R 2 of 0.74. These outcomes highlight the efficacy of strategic data collection, indicating its potential to enhance urban heat resilience and mitigate UHI effects.