Spectral nudging is an important method of improving the effectiveness of dynamical downscaling in the Weather Research and Forecasting model (WRF). However, spectral nudging is sensitive to nudging parameters, and most related research has typically adopted the default nudging schemes. Summer is the rainy season in China. Both typhoons and heavy rains occur in summer. It is therefore significant to discuss the optimal nudging parameter settings to achieve better summer climate simulations over mainland China. In this study, we analyzed the impacts of various nudged variables and their cutoff wavelengths on the downscaling ability by simulating the summer climate over mainland China in 2009 and 2010. The results showed that nudging the horizontal wind or potential temperature had a significant effect on the simulation of nearly all conventional meteorological fields and that nudging the geopotential height affected mainly precipitation. The suitable cutoff wavelengths were 500 or 1,000 km for wind, 1,000–2,000 km for potential temperature, and 1,000 km for geopotential height. Nudging all variables had obvious advantages over nudging the wind, potential temperature, or geopotential height individually in nearly all subregions. The appropriate cutoff wavelength was 500–1,500 km when all variables were nudged. Overall, nudging all three variables with 1,000 km cutoff wavelength is recommended for dynamical downscaling in summer over mainland China.