The invention discloses a comprehensive energy system situation awareness method jointly driven by model data, and relates to a comprehensive energy system state awareness technology, and the method comprises the following steps: 1, determining abnormal measurement based on an abnormal measurement detection method of a random matrix theory; 2, eliminating the abnormal measurement in the step 1, reading a line connection relation, and generating an impedance matrix of the integrated energy system; executing a situation understanding method of the comprehensive energy system to realize accurateperception of measurement of the comprehensive energy system; and 3, predicting the active power value of the power grid and the flow of the heat supply network pipeline by using an ultra-short-term prediction method based on a PID control strategy, executing the situation understanding method in the step 2 again, and estimating the operation state of the comprehensive energy system in the next time period. The method has higher detection precision and robustness, dynamic characteristics of the model and static characteristics of the model can be considered at the same time, and the ultra-short-term prediction method provided based on the PID control strategy has higher dynamic tracking capacity.