<p>Misinformation on social media has emerged as a modern weapon of warfare, disrupting societal peace, trust, justice, and democracy. It is quite challenging to address the issue of information credibility for microblogs. It becomes more challenging when the authenticity of the poster is hidden. The concept of information credibility has multi-perspectives. There are many necessary aspects of information credibility which must be considered for effective credibility assessment. It is observed that some important aspects of credibility are not considered in existing studies. The complete credibility assessment solution needs a comprehensive and diverse set of features for such complex identification. Therefore, these features are identified and proposed by exploring the related research studies consisting of the necessary credibility aspects. These features consist of diverse levels provided by microblogs. These levels include the post, poster, poster’s social network, and actual information propagation network. An exploratory study is also conducted to propose the best credibility features that are used in the proposed solution. The attempt is made for a hybrid features fusion model which combines feature-based or machine learning and graph-based approaches. It is a lightweight, high-performing, non-latent features model to avoid their drawbacks. It assesses the levels of credibility of the concerned post. It is designed for high-impact applications to combat low-credibility content during elections, crises, and other critical scenarios. The model is executed over a publicly available dataset extended for credibility assessment. The model provides good results with 95.6% accuracy by XGBoost using platinum features. The performance of the proposed model is compared with state-of-the-art that produced much-appreciating results.</p>