Abstract: Forest fire risk analysis and forest fire risk zoning are important parts of the forest fire management. MODIS burn scars of remote sensing data sets MCD45A1 of Heilongjiang Province in 2000-2010 was used to build the spatial logistic forest fire risk model based on the spatial distribution of forest fire and forest fire impact factor by using geographic information system technology. Forest fire risk zoning study was conducted in a larger temporal scale and provincial spatial scale. Logistic model of forest fire risk built by spatial sampling between the distribution of forest fires and forest fire impact factor fit well (p<0.05). The relative operating characteristic value was 0.753 and the probability distribution map of forest fire was gotten by layer computing. Forest fire area of Heilongjiang province was divided into none, low, moderate, high, and extremely high fire risk zones. Extremely high and high fire risk zone were located at Great Xing’an Mountain, while high or moderate fire risk area at Xiaoxing’an Mountain basically. Small parts of the eastern mountain were in moderate fire risk, and other areas in low or none fire risk. Quantitative and positioning evaluation of the forest fire in Heilongjiang province provides scientific basis for the prevention of forest fire fighting and rescue work.