This article systematically and quantitatively explores the applicability of the International Reference Ionosphere IRI-2016 model in polar regions, using observation data of the ionospheric F2-layer peak electron density (N_mF₂) acquired one solar cycle at three stations within the polar cusp region and the latitude of the aurora belt. The results showed good consistency between the IRI-2016 predictions and observation at the auroral latitude Tromsø Station, with relative errors less than 40% in most seasons. The accuracy of IRI-2016 prediction was marginally higher during solar maximum years than during solar minimum years at Tromsø, conversely it was higher during solar minimum years than during solar maximum years at the Zhongshan and Longyearbyen stations. At Zhongshan and Longyearbyen stations, the relative error is less than 20% in only a few months, more than 40% in most months, and is almost 100% in winter. Especially at Longyearbyen Station, the relative error is more than 100% in winter during solar maximum years. Overall consistency between observations and IRI-2016 prediction perform was lowest in winter, and highest in summer at all three stations. Plasma convection strongly influences NmF2 and particle precipitation in the polar regions, but the IRI-2016 model could not accurately reproduce these physical processes.