Various factors, such as solar radiation, the tilt angle of solar photovoltaic modules, and the conversion efficiency of these modules, play crucial roles in influencing the power output of photovoltaic (PV) systems. When it comes to predicting faults in distributed photovoltaic power sources, the primary focus and challenges revolve around the constraints imposed by maximum power. Consequently, a fault prediction method was introduced to distributed photovoltaic power sources, centered on the calculation of maximum admissible power. The methodology involved constructing an output model for distributed photovoltaic power sources to ascertain the limits of reactive power. By calculating the distribution of harmonic currents based on the maximum admissible power, the short-circuit capacity was determined. Faults in distributed photovoltaic power sources and categorized them into different fault warning levels. Through experimental analysis, it is demonstrated that the proposed method predicts photovoltaic power loss data more accurately, thus facilitating effective fault detection. Furthermore, the method computes the correlation of warning indicators for each node in the distributed photovoltaic power source, thereby establishing the warning fault status for each node. In the experiments, three nodes are in the green warning state, six in the yellow warning state, two in the orange warning state, and one in the red warning state, underscoring the favorable fault warning effectiveness of the proposed method.