The invention discloses a forest fire 
fire point detection method based on an MODIS. Firstly, 
time sequence data is established for 
multiple days of MOD02 and MOD03 data from the MODIS, 
time sequence daily value data of 
multiple days of NDVI and 
brightness temperature is obtained, break points in the 
time sequence daily value data of the NDVI and 
brightness temperature are recognized through a time sequence 
break point recognition algorithm, and the recognized break points serve as potential fire points; relative fire points are distinguished from the potential fire points through a space context 
analysis method; real fire points are extracted from autonomous modular sensor data, the relative fire points are compared with the real fire points, and the 
correctness and 
leakage rate of the relative fire points are verified. The potential fire points are distinguished through the method that the time sequence data break points are detected through the BFAST 
algorithm without determining the potential fire points through fixed threshold values, therefore the 
algorithm has higher adaptability and flexibility for 
fire point distinguishing, the defects that the regional difference cannot be met by distinguishing the potential fire points through the fixed threshold values is overcome, and the 
fire point correctly discriminating rate is increased.