A fuzzy logic based multiobjective genetic algorithm (GA) is used to optimize micromechanical densification modelling parameters for warm isopressed beryllium powder. In addition to optimizing the 19 main parameters of the model with 17 objective functions (experimental data points), the GA provides a quantitative measure of the sensitivity of the model to each parameter, estimates the mean particle size of the powder, and determines the smoothing factors for the transition between stage 1 and stage 2 densification. While the GA does not provide a sensitivity analysis in the strictest sense, and is highly stochastic in nature, this method is reliable and reproducible in optimizing parameters given any size data set and determining the impact on the model of slight variations in each parameter.