Recent scientific and practical interest in recommender systems has developed. This field has made great strides, yet there are still many unexplored regions. This study examines web-based travel information to create customized vacation packages. The specific characteristics of travel data that distinguish travel packages from normal commodities present a significant challenge for recommendation algorithms. Vacation packages are examined to start this study. Next, the Tourist-Area-Season-Topic model will be created. TAST can depict travel arrangements and visits using different subject distributions. This suggests that visitor behavior and natural landscape elements like geographic areas and trip seasons affect retrieved subjects. After presenting the subject model representation, we provided a drink technique for generating curated recommendations for customized vacation packages. Tourist-relation-area-season topic (TRAST) is a suggested expansion of the TAST model to capture passenger links within a journey group. The TRAST model, TAST model, and drink suggestion approach are tested using trip package data. The experimental results suggest that the TAST model captures travel data's distinctive characteristics. When suggesting holiday packages, the cocktail technique outperforms typical recommendation strategies. Travel group creation can be assessed using the TRAST model, which considers visitor interactions