A systems-level model of the Westinghouse PDU coal gasifier has been constructed based entirely on the data gathered during testing of this gasifier. It represents a purely empirical attempt to correlate the published data by means of multiple linear regression analysis, leading to prediction equations for output variables as linear combinations of a number of input variables. The prediction equations so derived do not have any physical laws incorporated into them explicitly. They rely entirely on measurements of the physical processes going on in the gasifier-cyclone combination. However, the choice of relevant input variables was made with the help of some intuition about the variables which affect these physical processes, based on work performed at PSI to model coal gasifiers. It should also be noted that the data comes from engineering tests which were not necessarily designed to produce systematic variation of all variables, so the data may contain some biases about which variables are varied. The computer implementation of the model is very fast and simple. For each output variable a set of 11 coefficients are prescribed. To obtain a predicted value of that output variable, for a specified set of values of the 10 input variables, a linear combination (dot product) of the 10 input variables is formed, using the 11 coefficients. The use of this empirical model is restricted to predictions for the Westinghouse PDU coal gasifier (and associated cyclone) within the range of operating variables represented in the data base. In particular, all the data involve operation on Pittsburgh Seam Coal. However, there is a range of coal feed rates, and both air-blown and oxygen-blown runs were included.