Previous oil palm simulation models such as OPSIM and SIMPALM have relied on direct measurements of vegetative biomass and biomass production as a basis for predicting total biomass production and bunch yield. To overcome the limitations imposed by the need for extensive on site measurements, methods of simulating vegetative biomass production (VBP) are needed. This paper describes approaches adopted for this purpose in the mechanistic, process-orientated model, OPRODSIM, a new comprehensive model of oil palm dry matter production and development. Using the standard method, leaf area index (LAI) is first simulated and used to give total VBP from an empirically derived curve. The VBP is then partitioned between the main vegetative organs, fronds, trunk and roots, using either constant or age-dependent partition coefficients. In an alternative method, frond biomass production (FBP) is derived successively from frond emission rate, single frond dry weight and planting density while root and trunk biomass production (RBP and TBP, respectively) are obtained from typical growth curves produced using data from the literature. FBP, RBP and TBP are summed to give total VBP. Age-dependent partition coefficients for frond, trunk and root biomass production are then calculated from the data and can be used for partitioning the VBP derived by the standard method. A further option is to derive FBP as for the alternative method and then partition the remaining biomass production between roots and trunk. Model outputs using various options are compared and the effects of additional modifying factors on VBP explored. A sensitivity analysis showed specific leaf area, planting density, single frond dry weight and frond number per palm to have the greatest effect on VBP using the standard approach while for the alternative method, only frond production and planting density were found important. In simulation tests for three data sets, the model gave mean values of VBP that were within 3% to 8% of actual means using the standard method and 4% to 10% using the alternative method. The consequences of the various options for bunch yield are also described.