This study implements content-based filtering for a tourist recommendation system in Pekalongan Regency. The method utilizes the TF-IDF algorithm to measure the weight of tourist attraction categories and cosine similarity to assess the similarity between attractions based on their categories. The dataset, provided by the Pekalongan Regency Tourism Office, includes eight categories: nature, water, artificial, culture, hills, beaches, camping, and adventure. The system is designed through the calculation of TF-IDF and cosine similarity to generate relevant recommendations. The findings show that the system effectively provides recommendations aligned with user preferences by presenting a list of tourist attractions with relevant categories. This system assists tourists in discovering destinations that match their interests while supporting the promotion of tourism in Pekalongan Regency.