Soil carbon dioxide (CO2) emissions from the field of corn (Zea mays L.) play an important role in global warming. This study investigated temporal variability of soil CO2 fluxes (Rs) with soil temperature (Ts) and moisture (θ) and built DAYCENT models for predicting future impacts of climate changes on Rs using the measured high-frequency data. Rs trend was tested by Mann–Kendall and Sen Estimator. Predicted Rss under different climate scenarios were compared using Parallel-line Analysis. The findings indicated that daily Rs exponentially increased with Ts constrained by θ. During the θ of 27–31%, there was a strong exponential relationship between Rs and Ts, but the relationship was weaker for the θ of 38–41% and 22–26%. Soil environmental index (SEI, Ts × θ) significantly impacted Rs with linear regression Rs0.5 = 0.4599 + 0.002059 × SEI in 2008, 2009 and 2011. At the diurnal scale, there were different trends in Rss and relationships among Rs and Ts and θ in different years. Predicted yearly Rss, root Rss and corn yield in 2014–2049 increased with an increase in temperature scenarios, but the Rss significantly increased as temperature rose by 1°C or higher. Predicted yearly Rss, root Rss and yield reduced with precipitation scenario increase, and the root Rss and yield significantly diminished as precipitation increased by 15 and 30%. Predicted yearly Rs from cornfields had a significantly increasing trend. Future research is needed to explore methods for mitigating cornfield Rs and evaluating sensitivities of different cropland Rss to temperature changes.