Polymer electrolyte fuel cells (PEFCs) have attracted attention as one of the next-generation power sources that are environmentally friendly. However, the high manufacturer's costs, short service life, and insufficient cell performance are solved to realize widespread commercialization. A non-platinum group metal catalyst is relatively low in cost and highly durable. In particular, the oxides of group 4 and 5 elements were reported to show high oxygen reduction reaction (ORR) activity and durability. Their oxides are considered dominant candidates for the non-platinum group metal catalysts, although they have a fateful flaw that their ORR activity is clearly poorer than traditional platinum-supported carbon catalysts. It was reported that introducing oxygen vacancies and other transition metals into the crystal structure of 4- and 5-group metal oxides effectively enhanced the ORR activity. The active site of 4- and 5-group metal oxides is unclear despite it being well-known that the oxygen vacancies and dopant atoms are related to its ORR activity. It is now indicated that candidates of active site were the oxygen vacancies around the dopant atom and the crystal distortion. However, it is hard to separate their effects because both dopant atoms and oxygen vacancies distort the local crystal structure. On the other hand, identifying the ORR active site leads to obtaining the catalyst design with high ORR activity. We focused on the oxide nanosheet, the simplest structure formed by a six-coordinate octahedral, as the model electrode. The thickness of oxide nanosheets is about 1 nm, and forming a single layer on a substrate is relatively easy. Thus, it was expected that the electron conductivity was guaranteed enough by the tunnel effect. This study employed titanium oxide nanosheets (TiO2ns) as the model catalyst and niobium as a dopant to investigate the effect of dopant atoms on ORR activity. The ratio of Nb/Ti in the prepared model electrode was 0/1, 1/1, and 2/1. In addition, the oxygen vacancies were introduced into all prepared electrodes to study the impact of oxygen vacancies on ORR activity at each Nb/Ti ratio. The thickness of Nb-doped TiO2ns (Nb-TiO2ns) was about 1.2 nm for the 1/1 ratio and slightly thicker for the 0/1 ratio. In contrast, the thickness of the 2/1 ratio of Nb-TiO2ns was about two times thicker than the 1/1 ratio. The thicknesses of Nb-TiO2ns in each ratio were increased by the reduction treatment to introduce the oxygen vacancies. One reason for this thickness increase may be related to the phase transition into the anatase structure. ORR activity was evaluated from the on-set potential of ORR (E ORR) measured at the step scan with 3 minutes step width. E ORR was slightly elevated with Nb doping. The ion radii of Nb and Ti atoms are 3 pm different when it is assumed that both ions are six-coordinated and their valence is Nb5+ and Ti4+. These results showed that the crystal structure distortion probably has a slight but positive effect on ORR activity.After the reduction treatments, the E ORR of the 0/1 ratio of Nb-TiO2ns increased and was the same as that of the 2/1 ratio. The introduction of oxygen vacancies certainly enhanced the ORR activity of the 0/1 ratio of Nb-TiO2ns. The oxygen vacancies are thought to be necessary to enhance the ORR activity, although the reduction condition wasn’t optimized for the Nb-doped samples.From the above results, we consider that the active site is the oxygen vacancy site rather than the crystal distortion site at this time.