CO2 flooding is one of the key technical methods for low permeability reservoirs in the middle and late stages of development, which helps to accelerate the realization of carbon neutral vision while enhancing oil recovery. However, continuous CO2 flooding in low-permeability reservoirs will form dominant channels for gas channeling, resulting in ineffective circulation of injected gas and greatly affecting the development effect of gas flooding, so it is very important to establish a method for evaluating the degree of gas channeling in low-permeability reservoirs. Firstly, we establish a numerical simulation model considering the parameters of the target block, take continuous gas flooding as an example, and carry out sensitivity analysis on reservoir property parameters, reservoir fluid parameters and production and development parameters, so as to clarify the influence of each factor on the law of gas channeling. Secondly, the gas channeling indexes were selected, and the weights of each index were determined by analytic hierarchy process, and optimized based on Genetic Algorithm (GA), Wolf Pack Algorithm (WPA) and Fireworks Algorithm (FA) respectively. Finally, based on the actual reservoir CO2 flooding gas channeling degree, select the optimal weighting scheme, establish the fuzzy evaluation standard of CO2 gas channeling degree, and determine the evaluation standard of gas channeling degree. The study shows that the main controlling factors of CO2 gas channeling are, from strong to weak, fracture spacing, injection and extraction well distance, permeability ratio and injection rate, while layer thickness and reservoir matrix porosity have less influence on the extent of gas channeling. Wolf pack algorithm has the best optimization effect on the weights of the influencing factors, and the root mean square error is 0.051, which proves that the scheme has high accuracy. The evaluation value of gas channeling degree of the target block is 2.6274, which is moderate gas channeling. This study provides theoretical support for evaluating the degree of gas tampering in CO2 gas flooding process, and provides scientific basis for subsequent gas tampering management and injection and extraction program design.