In complex V2X communication scenarios, the communication quality is affected by communication channel factors. Based on the V2V 3D channel model, this paper optimizes the parameters of three channel parameters, namely, the number of reflection points M, reflection coefficient $\mu$, and Rice factor K, and generates and outputs six typical scene C-V2X channel models. The K-Means++ algorithm is used to obtain the main class data, take the main class data as input, and GA (genetic algorithm) is used to optimize the channel model to obtain the optimal parameters. To verify the effectiveness of the channel, we designed an internal and external field cross test verification method based on the C-V2X off-road test platform. Experimental results show that the packet loss rate and delay of the in-chamber test are slightly lower than those of the actual outfield test data, which verify that the six communication channel scenarios proposed in this paper have been effectively established.