The invention discloses a detection method of insulator surface water content based on hyperspectral technology. The detection method comprises the following steps: shooting insulator surfaces with known different surface water contents by utilizing a hyperspectral imager so as to obtain a corresponding hyperspectral image, performing black-white correction, multiplicative scatter correction and smooth denoising on the hyperspectral image to obtain a hyperspectral line Xi of the surface water content Si, and then obtaining a water content BP neural network prediction model; shooting to obtaina hyperspectral image of a to-be-detected insulator, performing black-white correction, multiplicative scatter correction and smooth denoising to obtain the hyperspectral line X0 of the to-be-detectedinsulator, inputting the hyperspectral line X0 of the to-be-detected insulator into a prediction model to obtain the surface content of the to-be-detected insulator. Through the method disclosed by the invention, the non-contact real-time detection can be performed on the field insulator under an electrified condition, the operation is simple and the realization is simple; and the method can provide reliable and accurate test evidence for the dampproof and moisture-resistant design, manufacturing and maintenance of the insulator.