Glass fiber-reinforced polyamide-6 (GFPA-6) was compounded using a HYPERKTX46 twin screw extruder with three different types of screw segments (rotor segments, kneading disc segments and turbine segments). Based on the neural network regression analysis, we investigated the effects of compounding conditions, such as screw configurations, screw rotation speed and production rate on mechanical properties, such as tensile, flexural and impact strength of GFPA-6. It was found that the mechanical properties of GFPA-6 changed in different ways with compounding conditions (screw rotation speed and production rate), depending on the type of screw segments. It was also found that the rotor segment exhibited good mechanical properties in wider ranges of compounding conditions than the kneading disc segment and the turbine segment, probably due to its mild and uniform mixing capability.