Alzheimer's disease dementia (ADD) and amnestic mild cognitive impairment (aMCI) are characterized by spatial navigation deficits in unfamiliar environments mainly due to early hippocampal impairment. Spatial navigation abilities in individuals with non-amnestic MCI (naMCI) are, however, less understood. We aimed to evaluate real-space navigation performance in individuals with naMCI and to compare it with performance of cognitively normal (CN) older adults, individuals with ADD and aMCI. In all, 126 participants: naMCI with mainly attentional and executive dysfunction (n=38), aMCI (n=40), mild ADD (n=25) and CN (n=23) underwent clinical and neuropsychological evaluation, MRI brain scan and examination of spatial navigation in the real-space human analog of the Morris Water Maze (hMWM). The hMWM is designed to assess two separate types of spatial navigation (body-centered [egocentric] and world-centered [allocentric]). The participants were asked to search for a hidden place using a start position and two orientation cues (first training – ego+allo task), start position (second – egocentric task), orientation cues (third – allocentric task) and to recall the position of a hidden place after 30 minutes using orientation cues (fourth – allocentric delayed task). The analyses of covariance controlled for age and years of education with post-hoc Sidak's correction were used. The ROC analysis was performed to assess the sensitivity and specificity of hMWM in discriminating the naMCI group from the aMCI and ADD groups. There were no differences between the naMCI and CN groups in any hMWM task (p≥.467). The naMCI group had more accurate performance than the ADD group in all hMWM tasks (p<.001). The naMCI group had more accurate performance than the aMCI group in the allocentric tasks (p<.001). The size of the area under the ROC curve for allocentric performance discriminating the naMCI group from the aMCI and ADD groups was ≥0.77 (p<.001) and the values of optimal sensitivity and specificity were ≥82% and 63%, respectively. Spatial navigation in naMCI is relatively spared compared to subjects with ADD and aMCI. Spatial navigation testing of allocentric navigation in the hMWM may reliably distinguish individuals with naMCI from individuals with aMCI and ADD, where hippocampal impairment plays a prominent role.