NOx (NO2 and NO) plays an important role in controlling stratospheric ozone. Understanding the change in stratospheric NOx and its global pattern is important for predicting future changes in ozone and the corresponding implications on the climate. Stratospheric NOx is mainly produced by the reaction of N2O with the photochemically produced O( 1 D) and, therefore, it is expected to vary with changes in solar UV irradiance during the solar cycle. Previous studies on this topic, often limited by the relatively short continuous data, show puzzling results. The effect of the 1991 Pinatubo eruption might have caused interference in the data analysis. In this study, we examine the NO2 vertical column density (VCD) data from the Network for the Detection of Atmospheric Composition Change (NDACC). Data collected at 16 stations with continuous longterm observations covering the most recent Solar Cycles 23 and 24 were analyzed. We found positive correlations between changes in NO2 VCD and solar Lyman-α over nine stations (mostly in the Northern Hemisphere) and negative correlations over three stations (mostly in the Southern Hemisphere). The other four stations do not show significant NO2 solar-cycle signal.The varying NO2 responses from one location to another are likely due to different geo-locations (latitude and altitude). In particular, two high-altitude stations show the strongest positive NO2 solar-cycle signals. Our 1D chemical-transport model calculations help explain the altitude dependence of NO2 response to the solar cycle. NO2 solar-cycle variability is suggested to play an important role controlling O3 at an altitude range from 20 km to near 60 km, while OH solarcycle variability controls O3 at 40 -90 km. While observations show both positive and negative NO2 responses to solar forcing, the 1D model predicts negative NO2 responses to solar UV changes throughout the middle atmosphere. 3D global model results suggest complex roles of