It was aimed to develop a valid and reliable scale to determine the mathematical problem solving anxiety of 4th grade primary school students. Relational survey model was used in the study. The sample of the study consisted of 206 primary 4th students studying in2 public schools of Afyonkarahisar centre. The sample selected by cluster sampling. It was decided that the scale should consist of six factors according to literature review. These are arising from understanding based anxiety, planning and execution of the plan based anxiety, arising from controlling the solution based anxiety, arising from self-efficiency based anxiety, arising from person based anxiety, and arising from the structure of the problem based anxiety. The content validity of Primary Problem Solving Anxiety Scale (PPSAS) was determined by the Lawshe coefficient. Confirmatory factor analysis (CFA) was performed to test the 6-factor 43-item structure for construct validity analysis. Finding show that second-order CFA, the arising from person based anxiety was extracted from the scale due to its low fit with the general anxiety structure. The theoretical structure that emerged as a result of CFA was found to be highly fit with its 5-factor 27-item form. The Croncbach α reliability coefficient for the whole PPSAS was found to be .91. For the sub-factors, .77 for anxiety arising from the understanding process, .79 for anxiety arising from the solution process, .78 for anxiety arising from the control process, .82 for anxiety arising from self-efficiency and .80 for anxiety arising from the structure of the problem. For criterion validity, the relationship between PPSAS and the problem solving test by Ulu (2017) was investigated on 337 students. A moderate negative correlation were observed between both the whole scale and its sub-dimensions with the problem solving test. As a result, PPSAS was found to be valid and reliable with its 5-factor 27-item structure.