This work offers a new framework to modify precision agriculture methods by combining dynamic soil categorization with pattern recognition-based anomaly detection. Conventional approaches of soil classification neglect the temporal fluctuations of soil parameters and cannot sufficiently detect abnormalities that greatly affect agricultural productivity. Through a dual-phase design that integrates adaptive incremental clustering with advanced pattern- based anomaly detection, the proposed system effectively addresses these constraints. An enhanced Auto-Incremental Clustering Algorithm (AICA-Plus) enables dynamic soil classification by adapting to new soil samples and real-time environmental changes. Complementing this, a Temporal Pattern Recognition System (TPRS) employs sophisticated sequence modeling to detect abnormal soil conditions based on temporal variations in soil parameters. From field- collected soil samples, including pH levels, moisture content, nutrient concentrations, electrical conductivity, and organic matter content, the integration employs multi-dimensional feature extraction. The proposed methodology achieves a 22% improvement over existing techniques, reaching 94.7% classification accuracy. Anomaly detection performance demonstrates 96.2% sensitivity for actual abnormalities and a 35% reduction in false positive rates. Furthermore, the technique enables proactive agricultural interventions by identifying critical patterns of soil degradation three to five days earlier than conventional monitoring methods. Measurable agricultural advantages, including an 18% increase in crop yields, a 25% decrease in fertilizer waste, and a 30% improvement in water use efficiency, came from the application. These findings provide farmers with data-driven insights for best decision-making and sustainable farming methods, therefore proving the practical feasibility of the suggested framework for large-scale precision agriculture uses.