The invention discloses a blade detection method based on blade load analysis, and belongs to the technical field of wind driven generators. The method comprises the following steps: 1, monitoring and acquiring vibration signal data of a fan blade through a sensor for detecting vibration of the fan blade, and converting the vibration signal data into modal data; 2, constructing the modal data in the step 1 into a model through a finite element analysis (FEA) tool, and adding an orthotropic material constitutive relation and geometrical characteristics of a plate-shell theory on the model to obtain a cantilever beam model comprising parameters of a fan blade load F and a stiffness matrix K; 3, conducting static load analysis on the cantilever beam model obtained in the step 2, and calculating the deformation u of the cantilever beam model, wherein u = {F}/[K], {F} represents comprehensive force borne by the blade and comprises self gravity and wind power, and [K] represents the stiffness matrix; and 4, conducting dynamic load analysis on the cantilever beam model obtained in the step 2, applying full constraint to the root of the blade, conducting modal analysis on the single blade through finite element software, obtaining the first six-order vibration mode and the inherent frequency of the blade, and obtaining the vibration mode node position of the blade fault position in the blade bending direction, so that the defect position and the defect type of the to-be-detected blade are obtained.