This paper proposes an artificial intelligence (AI) system to optimize solar power for autonomous 3D printing machines in smart factories in terms of sustainability. The system integrates Long Short-Term Memory (LSTM) models for solar power forecasting, reinforcement learning (RL) for job assignment, and genetic algorithms for scheduling jobs. The goal is maximum throughput and energy efficiency with a least amount of downtime. A 60-day pilot was conducted in a simulated smart factory with a 12 kW solar panel and five 3D printers. Outcomes are 32% power efficiency improvements, 28% increase in throughput, and 53% decrease in downtime compared with baselines. Real-time prediction (Mean Absolute Error: 10.8 W), adaptive allocation, and scheduling facilitate sustainable manufacturing. The paper suggests the framework supports significantly the integration of renewable energy into additive manufacturing and scalable solutions for smart manufacturing.