Abstract: With the large-scale integration of new energy sources represented by photovoltaics into the grid, the randomness, intermittency, and volatility of their output pose significant challenges to grid planning and operation, adding complexity and uncertainty to grid planning. Traditional methods based on analyzing typical daily photovoltaic output characteristics are insufficient to accurately reflect the probabilistic distribution features of photovoltaic output, especially under the context of modern power systems, which require power balance across multiple time scales and scenarios. To address this, this paper adopts a confidence-based method to systematically analyze the output characteristics of photovoltaics in two typical scenarios: forward peak and reverse consumption. A photovoltaic power generation parameter database covering different seasons and confidence levels is constructed to provide a quantitative basis for grid planning, operation dispatch, and new energy accommodation, thereby improving the accuracy and adaptability of current active distribution network planning.