As Internet of Things (IoT) applications become more prevalent in our daily lives, security concerns in the IoT have received more attention from researchers. However, the existing IoT systems do not adequately meet the increased security requirements created by such vulnerabilities. In order to solve the issues with current techniques, this article proposes a lightweight authentication protocol based on the Physical Unclonable Function (PUF) chip. Our model could send actual data from genuine sensor nodes and become robust enough to resist many attacks. Fuzzy logic categorizes sensor nodes after at least one transaction with the server so that it does not authenticate the same nodes through all the phases again in the next interaction. The model has been proven to be secure and more efficient, has lighter computational ability, has reduced communication costs, and has less storage requirements. Many lightweight algorithms cannot incorporate both security and efficiency. This work has shown how the two can be saddled in a lightweight protocol. Furthermore, many attacks were addressed within this model, such as Denial of Service (DoS), replay attacks, eavesdropping, machine learning attacks, and side-channel attacks. Security verification and performance evaluations proved the effectiveness of our model.