The exponential growth of Internet of Things (IoT) devices has highlighted critical vulnerabilities in connected systems, creating opportunities for cybercriminals to exploit these weaknesses through botnets and other attack methods. This paper presents an innovative security architecture for IoT networks, using a hybrid honeynet to simulate real-world vulnerabilities in a smart greenhouse environment. By deploying honeypots such as Drupot, Amun, and Dionaea, the system captures and analyzes attack attempts, providing insights into intrusion tactics and enabling proactive countermeasures. Results from a 30-day observation period demonstrate the effectiveness of the proposed solution in detecting SQL injections, Denial of Service (DoS) attacks, and port scanning. The findings emphasize the importance of combining low- and medium-interaction honeypots to strengthen IoT defenses while maintaining scalability and resource efficiency. This study contributes to advancing IoT security by offering a robust and adaptable architecture to mitigate threats in increasingly complex digital environments.