Abstarct: In recent years, recommender systems have been used to determine tourist destinations. Due to the massive growth of the Internet, recommender systems can create a huge opportunity for tourism-related businesses. Travel recommender systems have challenges that by solving these problems, better accuracy can be achieved in recommender systems. According to previous researches, recommender systems can use the clustering approach for the recommender system. The usual clustering methods have problems that can affect the accuracy of recommender systems. In addition, providing travel plans is one of the needs of users. Also, in travel recommender systems, external knowledge that leads to the right offer is necessary. In this research, a travel recommendation system is designed, the input of which is data set information and external knowledge, and the output of which is travel suggestions to the user along with intercity scheduling. For this purpose, the clustering approach is used to categorize a set of features and offer to the user. On the other hand, we present a method for clustering that does not require output parameters and is created automatically according to the dataset of model input parameters. In addition, it has examined the density of the data and tried to improve the previous clustering challenges. Then, after clustering and choosing the right data set for the proposal, using genetic algorithm, it manages the travel plan of users so that users can improve the priority of their visits. By choosing the right genes from the features affecting the user's travel time and distance, we improve the selection options for the suggestion. In addition, by creating a knowledge baxse, we enrich the recommender system and influence the genetic algorithm process so that we can make the recommendation process more accurate. Also, we allow system administrators to apply policies in the process of suggesting the system so that they can manage the process of suggesting to users. Data from business trips have been used to evaluate the proposed system. Finally, by presenting the proposed method, the accuracy of the suggestions has increased at least two times compared to previous researches, and the possibility of travel planning has been created for users. In addition, the clustering evaluation was done and this method was able to automatically perform the clustering regardless of the external parameters and achieved balanced accuracy between the usual methods and solved the usual clustering challenges.