This study aims to develop an instrument that has never been created before, the statistically good character test. The instrument can be used to measure the character of Muslim students at the level of competence in knowledge, comprehension, and application. The test development applies a modified 4D model procedure into three stages, define, design, and develop. The reliability of the test was obtained based on the level of reliability, the level of difficulty of the questions, the discriminatory power, and the effectiveness of alternative distractors. The research was conducted in 2021 with a total of 50 students. The results of the study (1) reliability coefficient (α) 0.938, (2) the average item difficulty level (p) 0.643, (3) the average discriminatory power index (rpbis) was 0.522, (4) the alternative distractors were categorized as effective (50%) and ineffective (50%). Based on these findings, it shows that cognitive test devices are feasible to use and can be used to determine the character of Muslim students.