The Loess Plateau in northern Shaanxi is located in the ecologically vulnerable region of northwest China. In recent years, affected by the return of farmland to forestry (grassland), the regional vegetation cover has changed significantly, and the response relationship with meteorological elements has also exhibited complex spatial differentiation and uncertainty, which urgently needs to be clarified. Taking different geomorphic regions as response units, this study investigated the spatiotemporal evolution characteristics of the Normalized Difference Vegetation Index (NDVI) before and after the greening (1982-1992 and 2005-2015) of the Loess Plateau in northern Shaanxi, analyzed the differential response relationship between vegetation cover changes and meteorological elements, and explored the interaction effects and risks of meteorological elements changes on vegetation cover evolution through the optimal parameters-based geographical detector (OPGD) model. The results showed that: (1) the NDVI of the Loess Plateau in northern Shaanxi and each geomorphic subdivision presented a significant upward trend during the statistical period, with significantly phased differentiation. The trends in temperature and evapotranspiration in different geomorphic regions showed significant differences before and after the greening of the plateau, with a mainly non-significant increasing trend from 1982 to 1992 and a non-significant decreasing trend from 2005 to 2015. The spatial heterogeneity of precipitation was more significant during the two periods, with the northern region showing a continuously increasing trend, while the southern region reversed from a decreasing trend to an increasing trend. (2) The NDVI of different geomorphic regions was positively correlated with temperature and evapotranspiration, and the correlation between evapotranspiration and NDVI was stronger than that between temperature and NDVI. Compared with the pre-greening period, the response intensity of the NDVI to various meteorological elements increased from 2005 to 2015. (3) The OPGD factor detection results showed that a 5km spatial grid was the optimally spatial scale for evaluating the impact of climate change on vegetation cover change on the Loess Plateau in northern Shaanxi. The meteorological factors that had the greatest impact on the NDVI in different geomorphic regions were in order: temperature (Loess tableland), precipitation (Cover sand loess hills), evapotranspiration (Loess hilly mounds), precipitation (Loess beam hills), precipitation (Loess wide valley hills), evapotranspiration (Wind-sand hills), and precipitation (Soil-stone hills). The interactive detection showed that the interaction between the meteorological factors had a synergistic enhancement effect on the spatial differentiation of the NDVI, but the dominant interaction factors in different geomorphic regions differed slightly at different times. (4) The risk detection found significantly spatial differences in the influence of different evapotranspiration amounts on the NDVI. (5) The change in vegetation cover on the Loess Plateau in northern Shaanxi could not simply be attributed to ecological restoration measures such as the return of farmland to forestry (grass) but should be the result of a coupling of multiple factors, in which meteorological elements played a non-negligible role.