Traditional power system scenario generation often relies on simple statistical models or rule-based methods, neglecting the stochastic nature of renewable energy outputs and the correlation between wind and photovoltaic outputs. This leads to scenarios that fail to capture real system complexities. To address this, we propose a typical scenario generation method based on the Copula function and an improved ISODATA clustering algorithm. First, the Copula function constructs a multidimensional joint distribution model of wind and photovoltaic outputs to capture their dependency structure. Monte Carlo sampling then generates a large number of scenarios. To extract representative features and reduce data complexity, an improved ISODATA clustering algorithm is applied, incorporating dynamic parameter adjustment and split-merge operations for optimized clustering. Simulation results on real datasets show that this method generates more realistic scenarios while achieving a superior balance between scenario quantity and accuracy, supporting power system planning and optimization.