To address excessive randomness, low late-stage convergence efficiency, and premature convergence in the aphid optimization algorithm (AOA), this study proposes a soft-threshold aphid optimization algorithm (STAOA) from a search-dynamics regulation perspective. The soft-threshold function nonlinearly controls update amplitudes to adaptively suppress or release step sizes, enhancing local exploitation while preserving global exploration and achieving a dynamic balance between them. A soft-threshold-based perturbation strategy further improves the ability to escape local optima, forming a hierarchical search regulation framework. Experiments on 23 benchmark functions, the CEC2019 test suite, and agricultural unmanned aerial vehicle (UAV) path planning tasks show that the STAOA outperforms several representative metaheuristic algorithms in accuracy, stability, and convergence speed, verifying the effectiveness of the soft-threshold mechanism in search-dynamics regulation and UAV path planning optimization.