The widespread adoption of edge computing and Internet of Things (IoT) applications has significantly increased the demand for efficient task scheduling in resource-constrained environments. These environments often rely on low-power, cost-effective hardware, making it challenging to maintain optimal performance while managing computational workloads. Traditional load-balancing approaches, such as round robin and least-loaded strategies, frequently struggle to optimize resource allocation in distributed systems with limited processing power, such as Raspberry Pi clusters. These conventional methods often fail to adapt dynamically to fluctuating workloads, leading to inefficient resource utilization, increased response times, and potential system bottlenecks. To address these challenges, this research introduces a novel Resource Conscious Predictive Load Balancing (RCP-LB) framework, which leverages machine learning techniques to enhance task allocation efficiency. Specifically, the framework employs Support Vector Machines (SVM) to analyze real-time CPU and memory usage data, enabling the system to predict the optimal task allocation dynamically. By proactively distributing workloads based on resource availability, the proposed approach reduces latency, improves response times, and maximizes overall system performance. Comprehensive performance evaluations compare RCP-LB with traditional load-balancing techniques, demonstrating its superior efficiency, scalability, and adaptability in handling dynamic workloads. The experimental results indicate that the RCP-LB framework significantly enhances system responsiveness and resource efficiency, making it particularly well-suited for real-time IoT applications and edge computing environments. By providing a robust, intelligent, and resource-aware task scheduling mechanism, this research contributes to the advancement of distributed computing frameworks, offering a practical and effective solution for modern, re-source-constrained systems.