Introduction and aim: The appropriate sensorimotor control of shoulder is required for performing functions of upper extremity [, ] and can be identified as the ability to generate and modulate forces effectively during muscle contractions while matching a given force level . Nevertheless, the sensorimotor strategies of shoulder in chronic hemiparetic subjects remain unclear on the literature, because they are relevant to support neurorehabilitation management.Aims: The aim of the study was to evaluate shoulder sensorimotor control and maximal shoulder muscle strength during isometric abduction and flexion in chronic hemiparetic subjects.Patients/materials and methods: Thirteen individuals with chronic hemiparesis due to ischemic stroke and thirteen healthy subjects matched for gender and age. Peak torque (PT) and torque steadiness was measured during isometric abduction and flexion of shoulder at 45° of range of motion using Biodex System III dynamometer. Standard deviation (SD) and coefficient of variation (CV) were measured from the steadiness trial. Difference between dominant and non dominant limbs in both movements (flexion and abduction) was performed using Anova Two-Way. As no differences were found between dominant and non dominant limbs for all the variables, a pool of data was performed. Difference between control, paretic and nonparetic limbs for the SD and CV was performed using Kruskall Wallis Test following Mann-Whitney considering a significance level of 0.017, due to Bonferroni adjustment. For PT variable, Anova Two-Way, considering a significance level of 0.05, was performed.Results: Both paretic and non paretic limbs reduced PT compared to control (paretic - deficit for flexion: 50.27% and abduction: 56.56%; non paretic - deficit for flexion: 84.01%; and abduction: 71.49%). Paretic limb PT was lower than non paretic one (p < 0.05). No difference was found between the paretic and nonparetic limbs in both movements for all variables of steadiness (p > 0.05), but they were higher when compared to control limb (p < 0.05). During abduction, higher values of CV and SD were observed in paretic and non-paretic limbs than during shoulder's flexion.Discussion and conclusions: The sensorimotor control is bilaterally altered in hemiparetic individuals, according the movement performed. Thus, the results demonstrate the need to include interventions in the rehabilitation program including, for example, bilateral upper extremity exercises, muscle strengthening, exercises with visual feedback, and functional manual tasks training, as reaching movements.