This investigation considers a quantitative procedure for determining the Values of critical process parameters in melt spinning to optimize the qualities of denier, tenacity, breaking elongation, and denier variance in as-spun polypropylene yarn. An orthogonal array in the Taguchi method defines the minimum set of parameter-level combinations that are experimentally tested. The significant process parameters, namely the third extruder barrel temperature, spinning temperature, metering Pump speed, and take-up velocity, are identified on the basis of the analysis of variance and F test. After a confirmation experiment is conducted to ensure the reproducibility of the experimental results, the back-propagation neural network establishes a continuous system linking 10 process parameters and four qualities. The technique for order preference by similarity to an ideal solution can be used to obtain a performance measure for assessing multiple qualities. The genetic algorithm attempts to find parameter values for optimizing the quality performance, including the denier, tenacity, breaking elongation, and denier variance. Finally, the experimental results demonstrate that the smallest denier, largest tenacity, smallest breaking elongation, and second smallest denier variance of as-spun polypropylene yarn can be achieved with the proposed approach in melt spinning. (c) 2006 Wiley Periodicals, Inc.