This study uses Research and Development (RnD) method with the aim of looking at the assessment sheet that has existed so far, developing a prototype assessment of the Real Analysis subject, and testing its validity, reliability and effectiveness. Development of assessment prototypes refers to the Borg and Gall model. The assessment prototype consists of RMT ability test and a Real Analysis course assessment instrument designed to measure the subject cognitive domain with the RMT approach, and adjusted to the style of thinking according to gender. The technique of collecting data is by giving written tests for material in the real function line, followed by documentation, observation, interviews, and questionnaires. The initial product data analysis is validated by experts to obtain content validity. Furthermore, the instrument is tested on a limited basis to subjects who contracted 15 Real Analysis courses. The results of the trial assessment instrument with the RMT approach are measured by the product moment formula, while the reliability test uses alpha formula. The difficulty level of the test questions for the ability of the RMT and the distinguishing power are also measured. In the assessment instrument with the RMT approach adjusted to the subject's thinking style based on gender, the validity and reliability are measured. The results show that the Real Analysis course assessment instrument through a gender-based RMT approach that has been developed is valid, reliable, practical and effective.