A numerical algorithm is developed to simulate the injection-compression molding (ICM) process. A Hele-Shaw fluid-flow model combined with a modified control-volume/finite-element method is implemented to predict the melt-front advancement and the distributions of pressure, temperature, and flow velocity dynamically during the injection melt filling, compression melt filling, and postfilling stages of the entire process. Part volumetric shrinkage was then investigated by tracing the thermal-mechanical history of the polymer melt via a path display in the pressure-volume-temperature (PVT) diagram during the entire process. Influence of the process parameters including compression speed, switch time from injection to compression, compression stroke, and part thickness on part shrinkage were understood through simulations of a disk part. The simulated results were also compared with those required by conventional injection molding (CIM). It was found that ICM not only shows a significant effect on reducing part shrinkage but also provides much more uniform shrinkage within the whole part as compared with CIM. Although using a higher switch time, lower compression speed, and higher compression stroke may result in a lower molding pressure, however, they do not show an apparent effect on part shrinkage once the compression pressure is the same in the compression-holding stage. However, using a lower switch time, higher compression speed, and lower compression stroke under the same compression pressure in the postfilling stage will result in an improvement in shrinkage reduction due to the melt-temperature effect introduced in the end of the filling stage. (C) 2000 John Wiley & Sons, Inc.