Primary infection by Plasmopara viticola, the causal agent of grapevine downy mildew, is a critical event in viticulture, marking the onset of fungicide applications to protect grape production. This infection originates from the overwintering of oospores in vineyard soil and their maturation and germination under specific conditions. In Switzerland, the VitiMeteo-Plasmopara forecasting model is employed to calculate primary infection risk after the stage prediction of oospore maturity, when the sum of temperature degree-day (TDD, base 8 °C) from January 1 reaches a threshold of 140 °C. We conducted systematic oospore germination observations over three consecutive seasons (2022–2024) in Changins, Switzerland, complemented by a full dataset spanning 2004–2024, to identify the key climatic drivers of oospore maturity. Pearson correlation analysis identified strong positive associations between oospore minimal time of germination (MTG) and temperature-related variables (TDD, mean temperature, and vapour pressure deficit), while rainfall- and humidity-related variables (cumulative precipitation and the number of rainy days in the preceding 30 days) were negatively correlated with MTG. A principal component analysis confirmed this pattern, showing that mature oospores were more frequently associated with cooler, wetter conditions, whereas immature oospores occurred under warmer, drier conditions. Furthermore, BBCH stage 13 (three unfolded leaves) marked a clear shift in climatic drivers. Before BBCH 13, hydric variables and TDD were the main factors associated with shorter MTG. From BBCH 13 onward, TDD, mean temperature, and vapour pressure deficit became positively correlated with MTG, reflecting a longer germination time, thereby suggesting a shift in the pathogen’s germination dynamics. Based on pre-BBCH 13 drivers, we developed a generalised linear model that integrates cumulative precipitation beginning on January 1, the number of rainy days in the preceding 30 days, and TDD. Applied over 2004–2025, with a maturity threshold of MTG < 1.5 days, this generalised linear model produced maturity dates that were closely aligned with observed primary infections. When integrated into VitiMeteo-Plasmopara, it increased the proportion of correctly timed primary infection predictions from 75% to 85% over 20 years, corresponding to a 40% reduction in delayed forecasts, thereby supporting more precise fungicide timing and reducing unnecessary treatments.