Spine grape (Vitis davidii) is a major native grape species cultivated in south central China, and the industry is threatened by downy mildew outbreaks caused by Plasmopara viticola in this rainy and warm region. To examine and characterize the local epidemics of this disease in this region, experiments were conducted in Jingzhou in 2012–2013 with variety ‘Gaoshan-2’ and Zhongfang in 2013–2016 with variety ‘Xiangzhenzhu’ in Hunan Province. The number of diseased leaves with new infections (DL_Sum) and disease severity were assessed every 5 days. First, we used a mechanistic simulation model originally developed in Europe to simulate the primary inoculum seasons in these years and found that they ended in early May, suggesting little influences from oospores to the disease in summer. Then we used negative binomial regression to model the temporal changes of DL_Sum in summer. The predictive variables included weather variables and two disease risk indices derived based on two other empirical disease warning systems developed in Europe and North America. Relative humidity (RH) was identified the best predictive variable among all variables. Ordinal logistic regression identified mean temperature and days of heavy rainfall (>7 mm) as key weather factors for disease development. In general, RH was more important for infection process, and temperatures and rainfall were more important for symptom development and spore dispersal. Canopy structures could also have significant effects on disease development. These results may help for the development of new disease warning or forecast systems more suitable for this disease in southern China.