A chemometric approach was applied for determining quinolinic acid in human plasma by differential pulse polarography after solid phase extraction. A fractional factorial design was used to examine the significant experimental variables for the peak height maximization. A Doehlert design, which allowed a sequential response surface methodology to be performed, was applied to the variables scan rate and drop size. The results indicated that the scan rate had the greatest effect on the response peak height. The linear range was extended from 8.52 x 10(-8) to 1.34 x 10(-5) M and the limit of detection was 2.9 x 10(-8) M. The validation process consisted of a pre-validation study followed by the main validation in the plasma matrix. The robustness and the intermediate precision were evaluated by means of experimental design. A 3(4)//9 screening symmetric matrix and a central composite design were used to optimize the solid phase extraction procedure of the analyte from human plasma using anion exchange cartridges. The goal was to select the best retention, wash and elution solvents and their volumes in order to maximize the extraction efficiency using as the response the polarographic peak height. An extraction efficiency of 90% was found. The method was also applied to the determination of quinolinic acid in urine and the mean concentration in human plasma and urine, was found to be 3.7 x 10(-7) and 4.9 x 10(-5) M respectively.