Purpose After stroke, upper limb somatosensation can be impaired which affects motor control. Vision is often used to compensate for this. A clinical assessment which assesses the combined sensorimotor function in the absence of vision would be beneficial in studies targeting sensorimotor improvement. Methods and materials We adapted the Action Research Arm Test (ARAT) to be performed without vision and called this modified version the sensorimotor ARAT (sARAT). Sixty healthy participants and 22 participants with chronic stroke performed the ARAT, sARAT, Fugl-Meyer upper extremity assessment (FM-UE) and Erasmus modified Nottingham sensory assessment (EmNSA). Discriminative validity of sARAT was evaluated by comparing performance between healthy participants and participants with chronic stroke. Convergent validity was evaluated by correlating sARAT with FM-UE and EmNSA. Results Participants with stroke performed worse on the sARAT compared to healthy participants (p < 0.001), with median scores of 52 (IQR 32-57) and 57 (IQR 57-57), respectively. The sARAT showed high correlations with FM-UE and EmNSA (r = 0.80-0.90). Conclusions The sARAT shows good discriminative and convergent validity in people with chronic stroke. It allows a quick assessment of the combined upper limb sensorimotor function, by performing the well-known ARAT with the eyes closed.