Subsurface temperatures are substantially higher in urban areas than in surrounding rural environments; the result is a subsurface urban heat island (SUHI). SUHIs and their drivers have received attention in studies world-wide. In this study, a well-constrained data set of subsurface temperatures from Amsterdam, The Netherlands, is presented. The study demonstrates that, through modeling of centuries-long (from fourteenth to twenty-first century) urban development and climate change, along with the history of both the surface urban heat-island temperatures and ground surface temperatures, it is possible to simulate the development and present state of the Amsterdam SUHI. The results provide insight into the drivers of long-term SUHI development, which makes it possible to distinguish subterranean heat sources of more recent times that are localized drivers (such as geothermal energy systems, sewers, boiler basements, subway stations or district heating) from larger-scale drivers (mainly heat loss from buildings and raised ground-surface temperatures due to pavements). Because these findings have consequences for the assessment of the shallow geothermal potential of the SUHIs, it is proposed to distinguish between (1) a regional, long-term SUHI that has developed over centuries due to the larger-scale drivers, and (2) local anomalies caused by anthropogenic heat sources less than one century old.