Research on the quantitative remote sensing detection of the Rujigou coal fire area in Ningxia by using the night-time airborne hyperspectral imagery is reported in this paper. Based on synchronous measurement of spectra and temperature，the author performed the spectral analysis and information extraction of indicated ground objects and thermal anomaly for defining the right thermal band for temperature simulation，simulation formula and minimum anomalous temperature，with the detection precision attaining the 1∶2 000 scale. The coal fire area was delineated and the combustion intensity was detected accurately. The author also analyzed the thermal diffusion regularity and the corresponding relations between the remotely sensed thermal anomaly and the underground coal fire anomaly，and achieved the goal that remote sensing quantitative detection results could be directly applied to the design of fire-fighting engineering.