The Present Study aims at developing The California Critical Thinking Skills Test for University Students According to the Latent Traits Theory. To Achieve this aim, Scientific Steps are followed Starting with the Translating the Items and its Instructions from English into Arabic, then re-translated the Arabic version into English, taking into consideration the accuracy and comprehensiveness of the whole ideas included in the test. Several modifications are required by translation experts and which were considered later. Then the Arabic version of the test is exposed on an Arabic Language expert who recommended certain modification in the style of the Arabic translation, and which were all taken into accont. The final Arabic version of the test is exposed on a number of experts in the field of Educational and Psychological Sciences in order to check its suitable characteristics as well as those of the alternatives with respect to form and content. Consequently all the test items are considered logically suitable to achieve the aim of the present study. The test is applied on a pilot sample of (50) male and female students in order to ensure the clarity of the test items and instruction, and to estimate the time required to respond to the test and which is fond out to be (50) minutes. The test Has applied on a stratified random sample of (500) male and female university students. Rasch's unit- dimension model, which is one of the Latent Trait Theory models, is applied for item analysis. To verify model's assumptions, the follow steps are applied:1. One-dimensional: In order to verify this assumption, factor analysis is done by using (the Principal components Method). According, One factor meaningful to the test. The same factor depended on the lower limits of Guttman who believes that a factor is considered statistically significant when the (Eigen value) equals or more than (1), and depending on a ratio of (0.30 – up) to represent the items saturation ratio by the general factor according to Guilford's norm. No item is omitted from the test. another method is used to achieve One-dimensionality of the test which is by finding out Point Biserial correlation (i.e the relation of the item score with total score of the test) as an internal consistency index between the test items since it is directed to measure one trait or ability.2. Matching the items of the model depending on the (Chi² value) according to the outputs of the program. This Procedure shows that four Items are statistically significant on a level of significance of (0.05). Therefore, (18) individuals of the sample of the statistical analysis were eliminated who have unmatched responses with the model's assumptions. 3.The discrimination coefficient value is found out to be near (1), where the exact value is found out to be (1. 03).4.The independence of the scale to achieve objectivity of measurement is represent by Rasch's model, consequently, no items are removed.5.In order to get rid of fractions and negative signals (i.e. minus), the logget unit is changed in to the percentile WAT unit to find out the items difficulty and the individual's ability estimation. In the light of the statistical analysis above, suitable conclusions, recommendations, and suggestions for further studies are put forward.