Mathematics self-efficacy is an important personal attribute and self-belief that can influence students' learning and performance in the subject, as supported by the Social Cognitive Theory. Literature review has shown that due to the scarcity of research on primary school students' mathematics self-efficacy, there is a lack of validated instrument to measure this psychological construct in the local school context. This study sets out to fill the literature gap by examining the validity and reliability of the Mathematics Self-Efficacy Questionnaire (MSEQ) (14 items). The contents of the instrument were validated before the instrument was administered to 100 primary students. The sample size was recommended based on a 1:5 subject-item ratio. Findings from the Exploratory Factor Analysis (EFA) suggested that mathematics self-efficacy is a unidimensional construct. It is highly reliable and can be used to gauge primary school students' mathematics self-efficacy in a Malaysian school setting. The educational implications of this study are discussed in this paper.