The recent literature on land surface temperature (LST) and urban heat island (UHI) highlights the imperative for more studies on the relationship between LST/UHI and biophysical/socioeconomic factors, especially in arid environments. This study examines the spatio-temporal variations of LST or surface UHI (SUHI) induced by biophysical and socioeconomic factors in Riyadh, Saudi Arabia. Additionally, the normalization methods for comparing LSTs of different periods were examined. The LSTs of the study area for four years between 1985 and 2015 in June/July were derived from multi-date Landsat images. The SUHI index of the different land-use/land-cover types (high-density residential, medium-density residential, low-density residential, industrial, vegetation, and desert) was computed from the LST data to analyse their relationships. Thereafter, geographical random forest (GRF) analysis was used to determine the influence of biophysical and socioeconomic factors on LST/SUHI. The findings show differences in the minimum temperatures from 1995 in all the land-use types. The industrial area has the highest temperatures while the temperatures of the vegetation area are the lowest. However, the means of the normalised LST values depict decreasing values. The use of the normalized ratio scale (NRS) was not successful. The GRF analysis indicates that land use/land cover (65%) has the highest indirect influence on LST/SUHI index, followed by nighttime light (20%), traffic (16%), tweet density (15%), population (14%) and built-up area (4%). In conclusion, land use types and socioeconomic factors influence variations in LST/SUHI. The article contributes to the knowledge of planning for urban heat island mitigation by highlighting the influencing factors.