This dataset contains supplementary modelling outputs produced for the MSc thesis on daytime Surface Urban Heat Island (SUHI) analysis across Essen, Wuppertal and Soest, Germany. The files support the Random Forest modelling component of the study, which examined the relationship between ECOSTRESS-derived Land Surface Temperature (LST), SUHI intensity, urban morphometric indicators and environmental predictors during the summer periods of 2022 and 2024. The archive includes the full run-level Random Forest model inventory, retained model-configuration summaries, model-ready sample-size information, random test-set and spatial block cross-validation outputs, full variable-importance rankings, partial dependence plot outputs and related supplementary modelling files. These files provide the detailed modelling record behind the results reported in the thesis and are made available to support transparency, reproducibility and future reuse.