Abstract: Under the influence of large-area extreme weather conditions, the impact of distributed photovoltaic fluctuations on power system balance is significant and may lead to risks such as curtailment of solar power and load shedding. In response to these issues, this paper proposes a multi-level rolling warning method for distributed photovoltaic fluctuations based on interval analysis theory, aiming to provide a rolling warning of the potential harm of distributed photovoltaic fluctuations. Firstly, the power control mechanism for handling distributed photovoltaic fluctuations in the power system is clarified, and warning levels are established to determine the range of fluctuations that can be controlled by different power control measures, i.e., the warning thresholds corresponding to different warning levels. Secondly, based on the probability density of distributed photovoltaic fluctuations, the probabilities of each warning level are calculated by integrating the probability densities within each warning range. Finally, the differences in forecasting accuracy of photovoltaic fluctuations at different time scales are analyzed, and the rolling warning of distributed photovoltaic fluctuations is achieved by periodically adjusting the warning results. Case study results demonstrate that the proposed method can determine the thresholds for each warning range while providing warning results for different system operating conditions and photovoltaic fluctuation events. Moreover, the root mean square error of the warning results obtained with our method compared to those of the Monte Carlo method is only 1.6718%, thus verifying the effectiveness and applicability of the proposed method.