Unilateral spatial neglect (USN) is a highly prevalent post-stroke deficit. Currently, there is no gold standard USN assessment which encompasses the heterogeneity of this disorder and that is sensitive to detect mild deficits. Similarly, there is a limited number of high quality studies suggesting that conventional USN treatments are effective in improving functional outcomes and reducing disability. Virtual reality (VR) provides enhanced methods for USN assessment and treatment. To establish best-practice recommendations with respect to its use, it is necessary to appraise the existing evidence. This systematic review aimed to identify and appraise existing VR-based USN assessments; and to determine whether VR is more effective than conventional therapy. Assessment tools were critically appraised using standard criteria. The methodological quality of the treatment trials was rated by two authors. The level of evidence according to stage of recovery was determined. Findings were compiled into a VR-based USN Assessment and Treatment Toolkit (VR-ATT). Twenty-three studies were identified. The proposed VR tools augmented the conventional assessment strategies. However, most studies lacked analysis of psychometric properties. There is limited evidence that VR is more effective than conventional therapy in improving USN symptoms in patients with stroke. It was concluded that VR-ATT could facilitate identification and decision-making as to the appropriateness of VR-based USN assessments and treatments across the continuum of stroke care, but more evidence is required on treatment effectiveness.