The aim of this study was to provide a better understanding of how vegetation and building geometry influence the spatial distribution of air temperature and nocturnal cooling rates (CR) in a high‐latitude city. Intra‐urban thermal variations were analysed in two seasons (May–September and November–March) and in different weather conditions (clear, calm and cloudy, windy) in Gothenburg, Sweden. Simultaneous air temperature measurements were conducted for 2 years (2012–2013) at ten fixed park and street sites characterized by varying type and amount of vegetation, building geometry, openness and surface cover. Several spatial characteristics, including sky view factor (SVF) as well as the cover and volume of buildings and trees, were calculated within circular areas of radii ranging from 10 to 150 m. Spatial characteristics were found to explain air temperature distribution in the studied area to a large extent throughout the day and year, in both clear, calm as well as cloudy, windy conditions. The highest correlations were found for weighted calculation areas accounting for the influence of both nearest (10 m) and wider (25–150 m) surroundings. Park sites remained cooler than built‐up areas, with the most pronounced cooling effect (0.8 °C) on clear, calm days of the warm season. The most important factor governing CR around sunset was SVF. However, on clear, calm nights of the warm season, they were also enhanced by vegetation, indicating the influence of evapotranspiration. Minimum night‐time air temperature was governed mostly by the presence of buildings. Within the street canyon, a daytime cooling and night‐time warming effect of a street tree was observed, particularly in the warm season. The study shows the importance of various spatial characteristics describing openness, amount of vegetation and building geometry in analysing intra‐urban variations in daytime and night‐time air temperature.