In order to improve the effectiveness of tourist routes recommended results, and reduce the influence of interference data on the recommendation results, a proposal has been made of Tukey skeptical model based on M estimation of collaborative recommendation algorithm for tourist routes. Firstly, a framework of scenic route recommendation algorithm is to be constructed based on interest preferences of tourists, and a reliable neighbor model is to be constructed based on k-distance skepticism. Secondly, according to the proposed model, we propose a decomposition algorithm of matrix estimation based on robust Tukey M, thus constructing the user feature matrix and project feature matrix. By adjusting the similarity between tourists, the influence has been reduced of interference configuration items on the robust estimation of the characteristic matrix. Finally, the simulation test on the web crawl data set shows that the proposed algorithm is characterized with such advantages as with a higher overall tourist satisfaction, lower tourist pains, and better diversity of tourist attractions.