The invention discloses a method for tracking the maximum power point of a photovoltaic cell on the basis of fuzzy probability. The tracking method comprises the steps that N sampling points [ui, P(ui)] are obtained by taking epsilon as a sampling interval, wherein i is a positive integer which is smaller than or equal to N, epsilon ranges from 0.05 UOC / Ns to 0.5 UOC / Ns, UOC represents the open-circuit voltage of the photovoltaic cell, and Ns represents the serial number of the photovoltaic cell; probability functions Pro(i) are solved by constructing a spread function fD and a subordinating degree function fM, results of the probability functions Pro(i) are ordered from large to small, and union sets of Xi corresponding to front probabilities are sequentially selected as the searching range of the maximum power point, so that the sum of the front probability functions Pro(i) is larger than or equal to a probability threshold value delta; the maximum power point[uMPP, P(uMPP)] is solved in the searching range of the maximum power point. According to the method, the defect that close sampling is needed in a global scanning method under the multi-extremum condition is overcome, and the advantages of being high in tracking speed and good in environmental adaptability are achieved.