Urban heat island (UHI) amplification poses growing risks to public health and energy demand, yet long-term, city-scale diagnostics remain scarce for rapidly transforming metropolises. Here we present the first 40-year reconstruction of Moscow’s surface UHI (SUHI) using 125 cloud-screened summer Landsat scenes (1984–2024, 30 m). Land surface temperature was retrieved with a single-channel algorithm constrained by MERRA-2 water vapor fields and emissivity from NDVI-based method; accuracy is ±2◦C. SUHI intensity aggregated into annual and eight 5-year epochs. Eight Local Climate Zones (LCZs) were mapped from very high-resolution imagery and spectral indices to analyse morphology–temperature links. <br />The results show that the thermally significant SUHI footprint (> 2◦C) expanded by 35 % (≈ 1040 km2) and now covers just under 4000 km2. Growth was highly uneven: New Moscow added 125 %, Old Moscow 19 %, and the suburban belt 35 %. LCZ trends reveal the fastest warming in new development (+0.19◦C yr−1) and industrial areas (+0.13◦C yr−1); forest and landscape parks warmed least (< 0.06◦C yr−1). Built-up index correlates strongly with mean LST (r = 0.68). Scenario mapping, informed by official land use plans, projects further intensification of SUHI along the Kaluzhskoye and Kievskoye highways and adjacent streets by 2035. <br />The openly available Google Earth Engine & Python implemented workflow is transferable to other cities, providing a template for climate-resilient planning and targeted green infrastructure interventions.