The purpose of this study is to develop instruments of critical thinking skills that meet the validity, reliability, and criteria for good items. This research includes development research which consists of the design and implementation stages. The analysis in the study used quantitative analysis with the Confirmatory Factor Analysis (CFA) Technique. The results showed that the value of t value> 1.96 (a <0.05) which means that each significant item measures critical thinking skills. The value of person reliability of 0.75 and item reliability of 0.94 means the consistency of the answers from the good subject. The quality of the items was analyzed using the QUEST program, indicating that the index estimates of the difficulty of the items ranged from -1.65 to 1.55 means that there is no problem that is too difficult or too easy, and is classified as a good item. MNSQ's INFIT value ranges from 0.86 - 1.14 which means that the items match the Rasch model. Outfit value for each item ranged from -1.2 - 1.7 (t = 2.00) which means that all items were received.