The application of AI-driven microclimate control is vital to enhancing agricultural output and minimizing resource wastage. This paper aims to present how real-time IoT sensor data, machine learning, and mobile edge computing can be leveraged to improve crop yield prediction and promote sustainable farming practices. The system developed in this paper achieves 98% accuracy in field condition classification, providing valuable insights for farmers while addressing key challenges in the agricultural sector, such as climate variability, resource scarcity, and precision agriculture. It incorporates machine learning algorithms, real-time sensor data, and a user-friendly mobile application that enables farmers to manage their crops efficiently, predict yields, and optimize resource allocation. Some of the major features include role-based access control, predictive analytics, and offline functionality, ensuring usability even in low-connectivity environments. The integration of interactive data visualizations enhances real-time decision-making by presenting trends and predictions in an intuitive format. Using the Random Forest model, this system can label field conditions with an accuracy of 98%, significantly improving crop health monitoring and reducing prediction errors. Additionally, the model leverages precision agriculture techniques to provide tailored recommendations based on environmental conditions. Some challenges include sensor noise and computational complexity in certain models, which are mitigated through data preprocessing techniques. This work directly addresses inefficiencies in agriculture and contributes to reducing environmental degradation through the application of AI-based microclimate management. Future improvements will focus on integrating real-time IoT sensors, expanding crop variety support, and enhancing mobile edge computing capabilities for more robust and scalable precision agriculture solutions.