Urban surface warming is often investigated using satellite-derived land surface temperature (LST), but short multi-year time series are difficult to interpret because of interannual variability, limited cloud-free observations, and heterogeneous surface characteristics. This study proposes a practical framework for assessing the temporal consistency and multi-indicator coherence of intra-urban thermal patterns using Landsat observations. The analysis was conducted for Zagreb using Landsat 8/9 Collection 2 Level 2 data for the summers of 2015–2024. Annual composites of LST, the Normalized Difference Vegetation Index (NDVI), and the Normalized Difference Built-up Index (NDBI) were aggregated to 218 local committees as intra-urban spatial units. LST trends were estimated using linear regression, Sen’s slope, and the Mann–Kendall test, while changes in surface characteristics were analysed by comparing the initial and final periods. Results show a widespread increase in summer LST across Zagreb, with an average linear trend of approximately 0.21 °C yr⁻¹ and maximum Sen’s slope values of about 0.35–0.36 °C yr⁻¹. Although the statistical significance of the Mann–Kendall test was limited by the ten-year time series, 216 of 218 local committees showed positive ΔLST values, with an average increase of 1.80 °C. ΔLST was negatively associated with ΔNDVI and positively associated with ΔNDBI. The workflow integrates multi-year Landsat composites, cloud-based processing, and open tools for spatial and statistical analysis, providing a practical framework for interpreting temporally consistent intra-urban surface thermal patterns in complex urban environments.