Recommendation systems are sophisticated tools that help users navigate the web and receive personalized information tailored to their preferences. In today’s digital landscape, these systems are particularly prominent in the realm of e-commerce. The recommendation process plays a pivotal role in boosting profits and drawing in customers by offering targeted suggestions and enhancing the overall user experience. Recommendation systems play a crucial role in influencing the behavior of both sellers and consumers positively. These systems recommend accurate set of products or items to users based on their past ratings and interactions. The book recommendation system works similarly, making recommendations for books to users based on their past choices and interests to enhance their overall experience. Despite of several methods one problem, i.e., cold start problem persists. To resolve this issue, a hybrid matrix factorization technique combining Jaccard Similarity and Cosine Similarity is proposed. In this work, a real time Book Recommendation Dataset is used to conduct the experiment. To compare the performance of the proposed technique, few baseline algorithms are chosen. The comparison matrices such as precision, recall, and accuracy are used to analyze the performances. The experiment results show that the proposed algorithm outperforms the baselines algorithms.