This report describes development and application of a computational model for fluidized bed gasification of coal. A two phase bubbling bed reactor model accounts for the development of bubbles, clouds, and emulsion, with a new statistical treatment to predict bubble size. Coal char reactivity is described in detail, using models for carbon heterogeneous chemistry, pore structure, and mass transport previously developed at PSI. Volatile release is handled in a semi-empirical manner, and potential gas phase reactions between fuel and oxygen in the bubbles are allowed. The theory yields qualitatively correct trends in comparison to results from the U-GAS pilot plant. However, the data imply larger gas by-passing and particle carry-over than predicted, and the computed performance exceeds that observed. The computer program has been exercised to map out gasifier performance over a range of conditions. It is shown that proper fluidization often requires more gas flow than can be accommodated chemically, due to limited kinetics and mass transport. Increasing the bed temperature is beneficial, if not ruled out by coal caking and agglomeration tendencies. Increasing the pressure is not particularly advantageous, but major benefits in terms of off-gas heating value and gasifier efficiency can be obtained by reducing the coal feed size. Potential improvements to the theory are discussed.