The invention discloses a fault hanging wall and footwall fuzzy recognition algorithm based on probability statistics. The method comprises the following steps: calculating angle between two adjacent edges of a fault polygon in sequence; carrying out segmentation processing on the fault polygon according to the sizes of the angles to form a broken line segment set of the fault polygon; calculating weight of each upthrown side breakpoint, each downthrown side breakpoint and each stratum contact point of each broken line segment within a certain buffer area and distance between the breakpoints and the broken line segment; comparing the weights of different types of breakpoints of each broken line segment to determine whether fault attribute of each broken line segment is an upthrown side or a downthrown side; and for the broken line segments which cannot be judged based on weights, judging the fault attributes thereof based on the distance between the different types of breakpoints and the broken line segments. The method realizes intelligent fault hanging wall and footwall identification, so that not only efficiency is high, but also the precision problem of manual fault combination is solved, and fault hanging wall and footwall identification accuracy is greatly improved; and the method has high universality.