Texture, a critical property of Asian wheat noodles, is normally assessed by sensory evaluation. However, sensory evaluation may be impractical for wheat breeders and noodle researchers who need to evaluate a large number of samples and have limited sample. Instrumental Texture Profile Analysis (TPA) has been widely employed to evaluate Asian wheat noodle texture. Nevertheless, a standardized method for performing TPA on these products has not been established. A series of studies were conducted to develop a testing method to best relate TPA results to sensory texture characteristics of Asian wheat noodles. First, the optimum TPA testing conditions (crosshead speed and degree of deformation) were determined for each noodle category (alkaline, instant fried, salted flat, and salted round), and were defined as the conditions which best related their results to the sensory data. Partial Least Squares (PLS2) was used to examine relationships between sensory first-chew characteristics (hardness, cohesiveness, springiness, denseness, starch between teeth, and toothpull) and TPA output (peak areas and heights). Response Surface Methodology determined the optimum TPA conditions (crosshead speed and % deformation) as follow: 1 mm/s and 85 % for alkaline, 1 mm/s and 70% for instant fried and salted round, and 5 mm/s and 65 % for salted flat noodles. Second, the effects of two sample cooking factors: noodle weights (20, 50, 100 g) and noodle to water ratios (1:10, 1:20) and three holding factors: media (with, without water), temperatures (25, 55 °C), and times (2, 15, 30 min), on the TPA results were investigated. Cooking factors did not significantly affect the TPA results but higher holding temperatures, the use of water as a holding media, and longer holding time significantly decreased most TPA parameters' values. Third, relationships between TPA and sensory first-chew parameters were examined for each noodle category. Predictive models of each sensory first-chew attribute were developed using linear and nonlinear (Fechner and Stevens) models, with single and multiple parameters. Hardness could be satisfactorily predicted by a single TPA parameter (area 1 or area 2), but other attributes required multiple parameters in the models to be satisfactorily predicted. Different model types were selected for each sensory attribute and noodle category. TPA peak area 1 and 2 were the best predictors for first-chew characteristics of cooked Asian wheat noodles. / Graduation date: 1998