Industrial and mining land is an important heat source in mining cities, however this heat has not been separated from that of built-up land in most studies of the thermal environment. This study measures the response of land surface temperature (LST) to patch characteristics and the configurations of four sub-classified industrial and mining land parcels in Wu’an, a mining city in China. We use Landsat-8 and ZY03 satellite imagery to obtain the LST and land use type. A heating rate index (T_m) is defined as the increase in LST per 1 ha increase in patch area. Multivariate regression analyses is used to analyse the relationship between thermal indicators and landscape metrics, both at the patch and village scales. Our results indicate that, at the patch scale, the mean LST (LST_m) can be described well by nonlinear binary functions of the area and landscape shape index (LSI), with different forms for each type of industrial and mining land patch. When the LSI does not vary, LST_m increases with increasing area, within area thresholds (0–50 ha for smelter or processing lands, and 5–30 ha for mining lands), while for a constant area, the LST_m decreased with increasing LSI. Below the critical LSI (~2.2), more complex patch shape had lower T_m values. Above the critical LSI, T_m do not increase or decrease as LSI increased. At the village scale, patch configuration of industrial and mining land has an effect on (LST_v). The landscape percentage and aggregation index are both directly correlated with the LST of villages that contained industrial and mining land. Our findings show that the degree of LST_v response to a land patch configuration metrics change is, in order, iron and smelter land > coal processing land > coal mining land > iron ore mining land. This study provides an improved understanding of the thermal environmental response to mining activities, and can serve as a reference for future industrial layouts and land use planning in mining cities.