Rain-fed agriculture is highly susceptible to climatic conditions. Study on analyzing the trend of yield and water use efficiency of rain-fed maize under climate change conditions is useful for agricultural production planning, which can provide adaptation strategies to climate change for agricultural production. Two test sites of Heyang and Changwu in Weibei highland were set as the study area. Experiment data in maize field was used to evaluate the applicability of CERES-Maize. Finally, the meteorological data produced from regional climate model (RegCM4.0) was used to predict maize yield and its water footprint before 2050. Results showed that the CERES-Maize model could precisely simulate the rain-fed maize yield and phenology; the absolute relative error (ARE) of most samples was less than 10%; the CERES-Maize model had good applicability in the rain-fed agricultural area in Weibei highland. Using the CERES-Maize model to simulate the water footprint of maize production was more accurate and reliable than using traditional water footprint calculation method. Under the RCP2.6 climate scenario, maize yield showed an increasing trend with increasing temperature and effective precipitation during the growing season; under the RCP8.5 scenario, maize yield showed a declining trend with increasing temperature and decreasing effective precipitation during the growth period. The excessive temperature rise had an obvious negative effect on maize yield, and the precipitation was positively correlated with the water use efficiency of maize. In order to effectively deal with the negative impact of climate change on the rain-fed crop yields, measures should be adopted, such as reducing greenhouse gas emissions, enhancing soil water storage and preserving moisture capacity, developing rainwater catchment irrigation, and selecting and cultivating water-saving and drought-resistant varieties.