We studied the interactions of silica and titanium dioxide nanoparticles with phospholipid membranes and show how electrostatics plays an important role. For this, we systematically varied the charge density of both the membranes by changing their lipid composition and the oxide particles by changing the pH. For the silica nanoparticles, results from our recently presented fluorescence vesicle leakage assay are combined with data on particle adsorption onto supported lipid bilayers obtained by optical reflectometry. Because of the strong tendency of the TiO2 nanoparticles to aggregate, the interaction of these particles with the bilayer was studied only in the leakage assay. Self-consistent field (SCF) modeling was applied to interpret the results on a molecular level. At low charge densities of either the silica nanoparticles or the lipid bilayers, no electrostatic barrier to adsorption exists. However, the adsorption rate and adsorbed amounts drop with increasing (negative) charge densities on particles and membranes because of electric double-layer repulsion, which is confirmed by the effect of the ionic strength. SCF calculations show that charged particles change the structure of lipid bilayers by a reorientation of a fraction of the zwitterionic phosphatidylcholine (PC) headgroups. This explains the affinity of the silica particles for pure PC lipid layers, even at relatively high particle charge densities. Particle adsorption does not always lead to the disruption of the membrane integrity, as is clear from a comparison of the leakage and adsorption data for the silica particles. The attraction should be strong enough, and in line with this, we found that for positively charged TiO2 particles vesicle disruption increases with increasing negative charge density on the membranes. Our results may be extrapolated to a broader range of oxide nanoparticles and ultimately may be used for establishing more accurate nanoparticle toxicity assessments and drug-delivery systems.