The assessment of Islamic Religious Education (PAI) material that is a common problem in it is the use of monotonous assessments. Because the tools used are rote systems such as memorizing short letters, the names of the prophets, and so on. The lack of varied models of PAI questions, makes students accustomed to certain questions and not trained in critical thinking. Likewise, the process of creating questions and revising PAI evaluations has not optimally involved teachers, so the quality of the questions is less guaranteed. Based on the background above, this paper will discuss what evaluation principles should be applied to Religious Education materials at the elementary school level? What are the criteria for effective and efficient tests for PAI materials in elementary schools? And what is the method of analyzing PAI test items for elementary schools? And the objectives of this study are to determine the evaluation principles that must be applied to PAI materials at the elementary school level, to determine the criteria for effective and efficient tests for PAI materials in elementary schools, and to determine the method of analyzing PAI test items for elementary schools. This study uses a literature review method with a qualitative approach. Literature Review is an activity that focuses on a particular topic that is interesting to analyze in depth and critically against the content of the text being studied. The results of this study state that the concept of good evaluation consists of the principles of continuity, comprehensiveness, fairness and objectivity, using criteria references, the principle of integration, while evaluation of cognitive aspects at the elementary school level uses a formative approach, diverse instruments, context considerations, involving students in the evaluation process, developing clear assessment rubrics, involving parents, providing feedback, using technology, reviewing, continuity and improvement. With examples of essay tests, multiple choice, fill-in-the-blank, matching, and true-false tests. The affective domain and using observation techniques and the psychomotor domain with the formulation of indicators with operational verbs. And to produce good data analysis, namely with reliability, validity, discriminatory power, and level of difficulty.