In the realm of multi-layered network analysis, the complexities of node interactions and strategic dynamics pose significant challenges. Addressing these, our study introduces the Tripartite Evolutionary Game Strategy (TEGS), a novel integration with advanced machine learning techniques. This approach centers on the dynamic interplay among network nodes using diverse strategies, enhanced by machine learning to improve strategy effectiveness and network adaptability. Our experimental framework involved a 500-node scale-free network with 7.2% interconnectivity across layers. Key findings include an overall network cooperation rate of 62.2%, surpassing the traditional rate of 51.1%, and Average Earnings of 4.5 points, exceeding the usual 3.8 points. Strategy Stability was high, with Centrality Node Influence being significant. Network Adaptability was strong, markedly better than moderate levels in traditional methods. Machine learning algorithms demonstrated high performance, with 95.5% Accuracy, 93.1% Precision, 94.3% Recall, and an F1 Score of 93.5%. Training Time was efficiently confined to two hours, and Prediction Time was a rapid 50 ms, underscoring their applicability in real-time scenarios. These outcomes highlight TEGS combined with machine learning as a transformative force in network analysis and strategy optimization.