The aim of this study is to develop a framework for detecting fake news that integrates semantic analysis with a source credibility algorithm, aiming to improve precision in detecting fake news. The proposed system utilizes semantic analysis, named entity recognition (NER), and a newly proposed source credibility algorithm as its methods. The system under consideration is assessed using the Information Security and Object Technology Research Lab (ISOT) dataset, which comprises news stories. The proposed system is evaluated using metrics such as accuracy and precision. The obtained scores are compared with the scores of baseline models. The proposed approach achieves an accuracy of 99.56%, demonstrating near precision, recall, and F1 scores across news categories. Comparative studies indicate that this method surpasses existing fake news detection tools that rely on content-based filtering techniques. The results show that adding the source credibility assessment algorithm to semantic analysis and NER has improved news detection systems, making them much more accurate and reliable. The results highlight the importance of using natural language processing (NLP) techniques and credibility analysis of news sources in efforts to combat misinformation.