Maize is one of the most important crops in China, both as human food and livestock feed. However, drought has become the key climate disaster affecting the maize production recently, particularly in northern part of the country, a primary maize cultivation area in China. The temporal and spatial changes of drought impact on maize have therefore aroused more and more concerns. Most previous studies investigated the drought impacts through climate indices methods, such as SPI, which usually neglected the crop responses to different degrees of drought stresses and the different drought-resistant characteristics between crop genotypes. This study assessed the potential impacts of past drought anomaly (1961-2010) on yield of summer maize, using a process-based crop model, CERES-Maize. The model was drove with the daily weather data and soil data based in every 50 km ×50 km grid. The output of model included the rainfed yield and irrigated yield in grid, if and only if which grew summer maize. The yield difference between the irrigated and rainfed was defined as the potential yield loss caused by drought. And we used the anomaly percentage of potential yield loss to reflect the interannual and interdecadal variation of estimated impact of drought on summer maize yield. Then we calculated the location of the drought affecting center from 1961 to 2010,which were divided in 5 decades (1960s:1961-1970, 1970s:1971-1980, 1980s:1981-1990, 1990s:1991-2000, 2000s:2001-2010). At last, we examined the relationships between estimated yield impacts and recorded indices of atmosphere circulation, to understand the underlying drivers of past drought risks. The indices were averaged over the summer maize prior growing-season (Nov-Apr) and growing-season (May-Oct) respectively. Our results demonstrate: (1) the irrigated yield exhibited a slight and insignificant (P>0.05) decrease from 1961 to 2010, while the rainfed yield exhibited a slight and insignificant (P>0.05) increase. So the potential yield loss in summer maize due to past drought anomaly exhibited a slight and insignificant (P>0.05) decrease, suggesting a decreased drought risk for summer maize production in China. But the potential yield loss experienced a significant increase (P<0.05) in the past two decades (1960s, 1990s), indicating a temporal fluctuation of the risk. (2) The cultivation area with a largest estimated loss in production due to drought experienced a clear northeastern move during the past 50 years, because the production loss decreased in north China while increased in northeast China with the changes in precipitation and temperature patterns. Besides, the period of 1960s-1970s and 1990s-2000s had the largest drift of the drought affecting center. (3) Estimated yield loss and location of the main drought affecting area demonstrated significant (P<0.05) correlations to a few atmosphere indices related to North Pole Vortex and subtropical high system. A weaker North Pole Vortex before or during the maize growing season associated with a less serious drought effect with affecting center located in further northern and eastern areas, vice versa. Subtropical high system also affected the drought risk for summer maize, but differed with indices.