The invention discloses a cemented filling body mechanical response characteristic prediction method based on an SEM (scanning electron microscopy) image. The method includes the steps: firstly, making an SEM sample; secondly, forming an SEM image in a scanning manner, and storing the image into a computer; thirdly, performing Gaussian filtering processing on the SEM image; fourthly, acquiring a plurality of clustering images of a cemented filling body; fifthly, determining a microscopic pore image of the cemented filling body to obtain a microscopic pore binary image of the cemented filling body; sixthly, combining the SEM image after Gaussian filtering processing and the microscopic pore binary image of the cemented filling body to obtain a testing sample image; seventhly, performing normalization processing on the testing sample image; eighthly, inputting the normalized testing sample image into a pre-built Tensorflow deep learning mechanical response prediction network to obtain auniaxial mechanical response prediction result. The method is high in prediction efficiency and has important significance for research of strength and stability of the cemented filling body, and consumed manpower and material resources are less.