Objective: To investigate the disruption of white matter (WM) integrity in patients with subcortical vascular cognitive impairment (SVCI) with multimodal structural magnetic resonance imaging (MRI), and to explore the relationships between WM damage, structural network disruptions, and general cognitive function decline.Methods: Twenty-eight patients with SVCI and twenty control subjects underwent structural MRI scans and neuropsychological assessment. WM volume (WMV), WM hyperintensities (WMH), lacunar infarcts (LI), diffusion tensor imaging parameters and structural network characteristics were compared between two groups. Correlations between these parameters and general cognitive function were calculated. Results: WMV, WMH load, LI number, mean fractional anisotropy (FA), peak height and peak location of mean diffusivity (MD) of normal-appearing WM (NAWM) and structural network characteristics were significantly different between the two groups. Disruption of WM microstructure and network characteristics was widespread. WMV, WMH load, mean FA and peak location of MD of NAWM explained about 70% of the variance in structural network characteristics. Shortest path length with LI number explained 47.5% of the variance in z-scores. Conclusion: SVCI was associated with widespread disruptions of WM integrity in the brain. Network characteristics may be a comprehensive reflection of WM integrity and a superior predictor of general cognitive function, while LI was an independent predictor of cognitive impairment.