The study suggested, developing a recommendation system that improves YouTube ad relevance by matching advertisements to video content. By gathering data from six categories and using word embeddings with machine learning models, the system identifies relevant ads that align closely with the topics users are watching, enhancing their overall viewing experience. The study compares models like logistic regression and random forest, finding that content-based filtering reduces ad fatigue by providing more meaningful and engaging ad placements. This system offers a more user-friendly approach to online advertising and sets the stage for further improvements in ad personalization and targeting.