The invention discloses an EIT tomography-based different fluid model identification and pipe blocking positioning method, which comprises the following steps of: 1, applying electric excitation to afluid conveying pipeline, measuring to obtain a boundary voltage sequence, and processing the measured boundary voltage sequence by a computer to obtain an EIT image; 2, inputting the EIT image into aTensorflow fluid model identification network by the computer, and identifying a core flow model, a laminar flow model, a bubble flow model and a circulation model; 3, the computer calculates the minimum error value of each fluid model; 4, the computer calls a fluid model optimization module to optimize the fluid model to obtain an optimized reconstructed image of each fluid model; Fifthly, the blocked pipe blocks are positioned through a computer; and sixthly, the computer judges the pipe blocking degree. The pipe blocking device is novel and reasonable in design, capable of accurately positioning in real time and judging the pipe blocking degree, capable of improving the conveying efficiency and reducing the conveying cost, high in practicability and convenient to popularize and use.