Generally, after a power outage, because of the loss of load diversity, the temperature-controlled load will experience a surge in load demand within a short period of time when the power supply is restored. This will lead to slow load recovery and even secondary power outages. This kind of phenomenon is called cold load pick-up. Therefore, it has certain practical significance to accurately extract the temperature-controlled load and analyze the influencing factors of the cold load pick-up. First, in this paper, the non-temperature-controlled loads in spring and autumn obtained by load decomposition are used as the reference load for the whole year. Then we strip the temperature-controlled load from the summer load. Compared with the existing methods, the method in this paper has more accurate extraction results. At the same time, to explore the deep-level relationship between load and temperature, in addition to decomposing load, this paper also decomposes temperature. According to the extracted temperature-controlled load, we use the Monte Carlo sampling method, taking an air conditioner as an example, to analyze the influence of the power outage time and ambient temperature on the cold load pick-up. Finally, through the load and temperature data of the PJM power market and the IEEE33 node system, the validity and accuracy of the extraction method of temperature-controlled load and the analysis of the influencing factors of cold load pick-up proposed in this paper are verified. This work is supported by the National Natural Science Foundation of China (No. 51777058).