Background. The purpose of this study was to investigate the effects of upper extremity immobilization and consequent walking speed on spatiotemporal gait parameters in stroke patients with hemiparesis. Methods. The following variables were assessed or measured in 29 stroke patients: age, height, weight, disease duration, Korean version of the Mini-Mental State Examination (MMSE-K), Berg balance scale (BBS-K), functional gait assessment (FGA-K), cause of the disease (type of lesion), and hemiparetic side. The measurement of gait was performed using two pressure plates of 1.5 m to create a 3 m walking distance and leaving 1.5 m of extension at both start and end, to ultimately create a 6 m walking distance that the patient could walk through. The following gait patterns were randomly selected based on card draws: self-selected walk speed (SW), self-selected walk speed with immobilized upper extremities (SWI), fast walking (FW), and fast walking with immobilized upper extremities (FWI). Each patient was assessed for four different gait patterns, with three measurements per pattern (12 gait measurements in total). Results. While there were significant differences in the stride length, step width, velocity, and step length of the paretic side between self-selected walk speed (SW) and SWI, FWI did not show significant changes in any of the tested parameters. Conclusions. Immobilization of the upper extremities may affect walking at self-selected walk speeds. A comprehensive training program including upper extremity movement should be established for gait rehabilitation. Clinical Trial Registration. This trial is registered at http://cris.nih.go.kr/cris.