To address the problems of existing swarm intelligence optimization algorithms, such as limited search information structure, reliance on mean statistics for group collaboration, and susceptibility to premature convergence, this paper proposes a novel swarm intelligence optimization method—the Yangtze Finless Porpoise Optimization Algorithm. Inspired by the cooperative hunting behavior and multi-frequency echolocation mechanism of the Yangtze finless porpoise, this algorithm constructs a new optimization framework from three levels: information perception, group collaboration, and search direction control. The algorithm achieves multi-scale parallel search through a multi-frequency echo hierarchical search mechanism, introduces a non-mean sonar potential field to model group collaboration relationships, and designs a cognitive fatigue-driven strategy switching mechanism and a direction-reversible search model to enhance the algorithm's global exploration capability and local development stability. Through systematic mathematical modeling and mechanism analysis, this paper theoretically demonstrates the rationality and potential advantages of this algorithm in continuous optimization problems. This research provides a new approach for structural innovation in swarm intelligence optimization algorithms.