This paper presents a novel Random Time-Slotted Multiple Access Protocol with Load Balancing for Direct-to-Satellite Internet of Things (IoT) applications. As Direct-to-Satellite IoT continues to grow globally, scalability becomes a critical challenge, particularly concerning Low-Earth Orbit (LEO) satellites covering vast regions with numerous connected devices. Existing protocols like LoRaWAN face limitations in handling massive device densities, leading to frequent collisions and reduced throughput. Our proposed protocol introduces a modified Slotted Aloha approach, enhanced with load balancing and beacon synchronization. By dividing communication sessions into distinct access intervals tailored to different packet sizes, and employing randomized time-slot allocation, our solution aims to maximize throughput efficiency while reducing collision probabilities. Simulations demonstrate that the proposed protocol achieves near-optimal throughput comparable to theoretical Slotted Aloha limits, significantly outperforming traditional Pure Aloha approaches. Future directions include refining the model to incorporate radio signal propagation dynamics and implementing field tests using actual hardware.