<p>Critical mathematic thinking ability is very important to solve daily problems. But in reality, junior high school students’ critical  mathematic  thinking  ability  is  still  low.  Ability  measurement such  as  measurement of  critical mathematic thinking ability cannot be measured through multiple choices test. In that case, an essay test in which graded scoring is used as scoring technique more suitable than multiple choices test. The result of the essay test will be analyzed to describe the already tested ability. There are two approaches in the measurement analysis; classical test theory and item response theory (IRT). The classical test theory has some weaknesses because it only depends on how many the right answers student could achieved. Meanwhile, the IRT technique is more suitable to analyze ability because lies on the pattern of the response and parameter of item test. Graded response models (GRM) is one of the IRT models that analyzed graded response.</p><p>The purposes of this research are to know about the result of the item parameter estimation of the test which has been developed  by  the  researcher  and  to  know  the  result  of  student’s critical  mathematic thinking  ability  parameter estimation  through  GRM  (Graded  Response  Models).  The  research  is  a  descriptive  quantitative  research.  The population of this research are 8th grade students of MTs Al-Ishlah Bobos and of SMP N 1 Dukupuntang in the academic year of 2012/2013. Applying purposive sampling method this research took 140 students as a sample, from whom 70 students from MTs Al-Ishlah Bobos and 70 students from SMP N 1 Dukupuntang. Measurement theory used in this research is Item Response Theory (IRT) with the GRM model and the instrument used to collect data is critical mathematic thinking ability test paper.</p><p>The result of the item parameter estimation shows that in terms of the item discrimination all four items tested are less good, meanwhile in the terms of item difficulty the results vary. The first item of the test is considered to be easy, the second and the third item of the test is considered to be very difficult, and the last item of the test is considered just difficult. The result of the critical mathematic thinking ability parameter estimation shows that 4,2% of students have very high critical mathematic thinking ability, 16,4% have high critical mathematic thinking ability, 65,7% have mean critical mathematic thinking ability, 13,5% have low critical mathematic thinking ability and there is no single student with very low critical mathematic thinking ability.</p><p> </p><span style="text-decoration: underline;">Key words : critical mathematic thinking, item of the test parameter, ability parameter, IRT, GRM </span>