In order to ensure that electric power equipment can operate normally under various complex environmental conditions, a multi-source data fusion sensing system architecture for the supply chain of power grid equipment is constructed. Support vector machine is used to classify the data sets of power equipment under various complex environmental conditions, and the main component analysis technology is used to reduce the dimensionality of power signals. Multi-source sensors such as temperature, humidity, vibration, electromagnetic interference, etc., are used to collect information on the operating status of power equipment, and subjective scores are given by evaluators to transform qualitative experience into quantitative, so as to realise a comprehensive assessment of the environmental status of power equipment. The results show that at the 720th time point, the predicted value of this paper's method is 13,009 kW, and the gap with the real value of 12998 kW stays within a small range. The integrated diagnosis time is in the range of 0.18–0.19 s, and the integrated diagnosis takes less time. The range of the fused evaluated values of the temperature parameter is narrowed to 29.5–30.5 °C, the range of the fused evaluated values of the humidity parameter is narrowed to 60–62%, which is closer to the actual value of 60–70%, and the range of the fused evaluated values of the vibration parameter is 0.36–0.48 mm/s. The method in this paper achieves a comprehensive and accurate assessment of the environmental adaptability of power equipment, which has high application value in power systems.