This study aimed at a comparison of item parameter estimation between traditional paper-and-pencil-based and computer-based testing. In addition to traditional item parameter, item response theory was applied for item parameter estimation and estimation of subjects’ ability. Further research on the number of subjects and its influence to parameter estimation in terms of variance analysis was also provide. After further data collection and analysis, major findings of this study were summarized as follow: first of all, when the numbers of subjects among all forms were closed, parameter estimation to either paper-and-pencil-based or computer-based testing was quite consistent—a result which was also identical to many scholars’ studies. However, controlling of certain amount of subjects indeed offers a decisive contribution to the stability of item parameter estimation.