The presence of heavy metals in wastewater, such as Ni, Pb, Cd, V, Cr, and Cu, is a serious environmental problem. This kind of inorganic pollutant is not biodegradable for several years, and its harmful effect is cumulative. Recently, semiconductor nanomaterials based on metal oxides have gained interest due to their efficiency in the removal of heavy metals from contaminated water, by inducing photocatalytic ion reduction when they absorb light of the appropriate wavelength. The most commonly applied semiconductor oxides for these purposes are titanium oxide (TiO2), zinc oxide (ZnO), and binary nanomaterials composed of both types of oxides. The main purpose of this work is to critically analyse the existent literature concerning this topic focusing specially in the most important factors affecting the adsorption or photocatalytic capacities of this type of nanomaterials. In particular, photocatalytic activity is altered by various factors, such as proportion of polymorphs, synthesis method, surface area, concentration of defects and particle size, among others. After a survey of the actual literature, it was found that, although these metal oxides have low absorption capacity for visible light, it is possible to obtain an acceptable heavy metal reduction performance by sensitization with dyes, doping with metallic or nonmetallic atoms, introduction of defects, or the coupling of two or more semiconductors.