Smart agriculture integrates IoT devices, sensors and wireless communication to improve crop productivity and sustainable farming practices. In India, climate change has significantly affected major crops, resulting in unpredictable yields over recent decades. Accurate crop yield prediction before harvest is essential for effective planning and resource management. This work proposes an AI-driven Intrusion Detection System (IDS) for securing smart agriculture networks while enabling intelligent crop yield prediction. The system combines IoT-based agricultural data with machine learning techniques. A Random Forest algorithm is employed to predict crop yield with high accuracy. An interactive web-based platform is developed to provide a user-friendly interface for farmers and stakeholders. Standard datasets such as NSL-KDD are used to train and validate the IDS module. The IDS continuously monitors IoT network traffic to detect cyber threats. This ensures data integrity and reliable system operation. The integrated framework enhances network security and decision-making. It supports efficient resource allocation and farm planning. The proposed system improves overall agricultural productivity. It also strengthens cybersecurity in smart farming environments. The approach promotes sustainable and data-driven agriculture.