Objectives (i) to create a shortened version of the Action Research Arm Test scale, (ii) to investigate its psychometric properties compared to the original scale and (iii) to externally validate it within an independent cohort. Design Prospective longitudinal cohort study. Settings Two University Hospitals (France, Switzerland). Participants 47 patients with poststroke motor deficits of the upper limb coming from two different sites were included and divided into two cohorts (n = 22 for the construction cohort; n = 25 for the validation cohort). Main Measures We used the first cohort to build the Mini-ARAT by shortening the Action Research Arm Test scale on the basis of ceiling/floor effects and collinearity of the subscales. We studied its reliability, validity, and responsiveness and performed an external validation with the second cohort. Results The Mini-ARAT consisted of 2 subscales from the original Action Research Arm Test scale (Grip and Pinch). Internal consistency (alpha = 87) and inter-rater reliability (0.99, 95% CI: 0.98-0.99, p < 0.01) were good and similar to those of the Action Research Arm Test scale. The Minimal Clinically Important Difference of the Mini-ARAT was 9 points. The predictive validity in the construction and validation cohorts showed good correlation between the Mini-ARAT at baseline and the Fugl Meyer at 3 months (rho, 95% CI: 0.77, 0.49-0.90, p < 0.01, and 0.58, 0.19-0.96, p < 0.01). Conclusion The Mini-ARAT is a time-effective tool able to capture the dynamics of motor deficits with high reliability and consistency, providing excellent information about residual motor functions, which is critically important for clinical and research purposes.