Abstract In order to facilitate management of large scale gas miscible EOR projects, a software package called FAME (Forecast Analyser for Miscible EOR) has been developed based on an analytical approach. The primary role of the software is to provide pool-wide performance predictions from mechanistic displacement characteristics determined either by part-pattern numerical simulations, Of by preliminary history matching of field data. The notable features incorporated into the software are that:important flood parameters, such as pattern allocation factors, solvent distribution and sweep, are dependent on pool-wide fluid movement and pattern interactions, which in turn are dependent on the injection/production sequence of historical and planned flood operations,crucial physical mechanisms associated with the gas miscible EOR, such as gravity override, solvent and chase gas miscibility with reservoir oil, and miscible and immiscible displacement processes are adequately represented, andresults of mechanistic numerical simulations or history-matching past field data can be mimicked through variations of a small number of parameters. The software has impacted the management of the Judy Creek hydrocarbon miscible projects with the achievement of a successful history match and the provision of useful information related to flood performance. Introduction During the operation of large-scale oil recovery projects, especially in enhanced oil recovery projects such as hydrocarbon miscible floods, it is highly desirable to be able to assess the project performance. Factors of interest include pool-wide fluid movements, local fluid distribution, sweep efficiency and displacement characteristics, as well as the prediction of future performance under different operating scenarios. Numerical simulation is a powerful method to assess the compound effects of numerous crucial mechanisms involved in various enhanced oil recovery processes. However, numerical simulation models for enhanced oil recovery processes require a large number of simulation grid blocks and small rime steps in order to adequately represent the critical mechanisms. A typical example of such a mechanism is the gravity override of injected solvent in gas miscible processes which causes low sweep efficiencies of solvent as described by H. Stone(1). The solid curve in Figure 1 is the theoretical relationship between the vertical sweep efficiency and water to solvent injection ratio (WAG ratio) at the fixed total fluid rate based on the Jenkins' analytical Expression(2) derived from the Stone's vertical conformance theory. The triangles and circles are corresponding outputs of the Todd-Longstaff type numerical simulations(3) using the horizontal and vertical numbers of grids of 10 × 3 and 15 × 13, respectively. In these numerical simulations, the sweep efficiencies were determined from values of tertiary oil recovery normalized by the tertiary oil potential at the two pore volumes of solvent injection. It is apparent in Figure 1 that the number of grids of 10 × 3 is not enough and overestimating the sweep efficiency, whereas the 15 × 13 is sufficient for representing the mechanism of gravity override. In three-dimensional models, several thousand grid blocks typically are necessary for part-pattern (such as a quarter five-spot pattern) simulations, even with a uniform geology description.