In this paper, a tourism revitalization model of intangible cultural heritage is constructed, and firstly, feature factors are extracted and multivariate features are combined using methods such as kernel density estimation to provide support for reliable prediction. Then decompose the MLP network structure, i.e. multilayer perceptron, and use Sigmoid function as the activation function. The importance of the feature factors is calculated using the random forest algorithm, and the collected tourism data are mined for the relationship between intangible cultural heritage and tourism revitalization to obtain the residual mean square of out-of-bag data. Finally, the coupling coordination degree is used to derive the development of ICH and tourism revitalization. The experimental results showed that 78% of the tourists were moderately to highly involved in the interactive participation of ICH, and only 22% were not interested in it. Therefore, the combination of ICH tourism and new media may become a new way of communication, allowing more general public to have the opportunity to approach and know ICH tourism.