Unilateral Spatial Neglect (USN) is a neuropsychological syndrome characterized by the failure to detect or respond to stimuli in contralesional space, independent of primary sensory or motor deficits. It occurs most frequently after right hemisphere stroke, affecting up to 82% of patients, and manifests with considerable clinical polymorphism. Subtypes such as egocentric, allocentric, and representational neglect disrupt a wide range of daily activities, resulting in prolonged hospitalization and reduced autonomy. Traditional paper-and-pencil assessments, however, lack sensitivity and ecological validity. To overcome these limitations, virtual reality (VR) combined with eye-tracking (ET) has emerged as a promising approach to identify oculomotor markers of USN in realistic environments. In this study, we developed a new multi-task VR-ET battery by analyzing fixation, saccade, and smooth pursuit metrics in static and dynamic tests, and compared these with conventional behavioral measures. A cohort of 24 stroke patients and 29 healthy controls completed standard paper-and-pencil tests followed by a new nine-task VR-ET battery. ROC analyses revealed that oculomotor metrics—particularly number of saccades, refixations, saccadic distance, and cumulative fixation time—achieved excellent sensitivity and specificity (AUC > 0.9), consistently outperforming behavioral metrics. Importantly, a reduced version of the battery reached perfect classification (AUC = 1) and detected cases missed by traditional assessments, highlighting its immediate clinical utility. Overall, these findings demonstrate that the VR-ET battery provides highly sensitive, ecologically valid, and time-efficient diagnostic tools, paving the way for technology-based diagnosis and personalized rehabilitation strategies tailored to patient's oculomotor profile.