In order to deal with time delay characteristic as well as dynamic characteristic widely exist in industrial process systems

this paper puts forward a dynamic soft sensor method based on joint mutual information. In the proposed method

maximizing the joint mutual information is taken as the criterion for selecting the continuous sub-variable set from each auxiliary variable's historical input data matrix. The selected sub-variable set forms a new data set containing the process's time delay information and dynamic information

which can be used to determine the time delay and historical data length of each auxiliary variable. The determination of the parameters of each auxiliary variable involves only the process's historical data and thus is independent of the following establishment of soft sensor model. As a result

the algorithm for constructing dynamic soft sensor models can be freely chosen according to the nonlinear degree of the specific application. The proposed soft sensor is applied to real-life debutanizer column process and its effectiveness is verified by the simulation results.