A travel recommender system can generate suggested itineraries for users based on their preferences. However, current systems are not capable of simultaneously considering trip length, distance, user requirements and preferences when making recommendations, being only equipped to consider one or two of these variables at one time. Also, to generate recommendations the system must process all attractions in the database, requiring more data access and longer processing time. We analyzed the check-in records of users and utilized a new concept of time intervals combined with a multiple days trip algorithm to produce itineraries compatible with the interests and needs of users. By applying R-tree to the travel recommender system, we reduced data access times and computation time. Lastly, we propose a trip evaluator equation that can be used to compare the strengths and weaknesses of each algorithm. Experimental results verified the effectiveness of our method.