The invention discloses a forest wireless sensor network positioning method, comprising (1), dividing a whole positioning region into m positioning subregions according to dispersion coefficients of RSSI (Received Signal Strength Indication) in different areas and fitting established RSSI path loss model R of each positioning subregion to form an RSSI path loss model R of the whole positioning region; (2), fusing the RSSI path loss model R of the whole positioning region and a logarithm path loss model R of the whole positioning region, thereby obtaining a forest path loss model R; (3), determining a plurality of wireless sensor positioning nodes through adoption of a subregion trilateration positioning method according to the forest path loss model R; and (4), clustering the plurality of wireless sensor positioning nodes through adoption of a k-means clustering method and determining final wireless sensor positioning nodes. According to the method, influences of a complex forest environment on the strength of wireless sensor signals are taken into full consideration, and the positioning for wireless sensors can be obtained accurately.