Aimed at phenomena that human can effectively sense various complex information, sensor human-simulated based on information fusion modeling is discussed, and several modeling methods is analyzed. The paper takes area measurement of irregular graphics as the researched object, and builds the model of sensor human-simulated with powerful performance of nonlinear fitting of support vector machine (SVM). Simulation studies show that the model of sensor human-simulated is satisfactory at precision. Therefore, it is significant for sensing complex variables to build effective model of sensor human-simulated by means of capturing various features of complex object.