Previous studies found that post-stroke motor impairments are associated with damage to the lesioned corticospinal tract (CST) and hyperexcitability of the contralesional cortico-reticulospinal tract (CRST). This proof-of-concept study aims to develop a non-invasive brain stimulation protocol that facilitates the lesioned CST and inhibits the contralesional CRST to improve upper extremity rehabilitation in individuals with moderate-to-severe motor impairments post-stroke. Fourteen individuals (minimum 3 months post ischemic stroke) consented. Physician decision of the participants baseline assessment qualified eight to continue in a randomized, double-blind cross-over pilot trial (ClinicalTrials.gov Identifier: NCT05174949) with: (1) anodal high-definition transcranial direct stimulation (HD-tDCS) over the ipsilesional primary motor cortex (M1), (2) cathodal HD-tDCS over contralesional dorsal premotor cortex (PMd), (3) sham stimulation, with a two-week washout period in-between. Subject-specific MR images and computer simulation were used to guide HD-tDCS and verified by Transcranial Magnetic Stimulation (TMS) induced Motor Evoked Potential (MEP). The motor behavior outcome was evaluated by an Fugl-Meyer Upper Extremity score (primary outcome measure) and the excitability of the ipslesoinal CST and contralesional CRST was determined by the change of MEP latencies and amplitude (secondary outcome measures). The baseline ipsilesional M1 MEP latency and amplitude were correlated with FM-UE. FM-UE scores were improved post HD-tDCS, in comparison to sham stimulation. Both anodal and cathodal HD-tDCS reduced the latency of the ipsilesional M1 MEP. The contralesional PMd MEP disappeared/delayed after HD-tDCS. These results suggest that HD-tDCS could improve the function of the lesioned corticospinal tract and reduce the excitability of the contralesional cortico-reticulospinal tract, thus, improving motor function of the upper extremity in more severely impaired individuals.