Downy mildew is an important disease in grapevine, which causes large losses of grape commercial production. The aim of this work was to investigate non-invasive sensing technologies and artificial intelligence applications for assessing downy mildew in grapevine under laboratory and field conditions. Machine vision was applied to assess visual symptoms while hyperspectral imaging was used to explore its potential capability towards an early detection of this disease. Image analysis applied to RGB leaf disk images taken under laboratory conditions was used to estimate downy mildew severity in grapevine (Vitis vinifera L.). Under laboratory conditions, a determination coefficient (R2) of 0.76 and a root mean square error (RMSE) of 20.53% were observed in the correlation between downy mildew severity by computer vision and expert’s visual assessment. Furthermore, an accuracy of 81% was achieved for an early detection of downy mildew using hyperspectral images. The classification of images of healthy and infected leaves, assessed on RGB images taken under field conditions yielded an accuracy of 89% using machine learning and deep learning techniques. These results indicated that non-invasive sensing technologies and computer vision can be applied for assessing and quantify visual symptoms of downy mildew in grapevine leaves. Severity of this key grapevine disease can be evaluated under laboratory conditions, but also under field conditions in commercial vineyards. In conclusion, computer vision, machine learning, and deep learning could be applied for key disease detection in grapevine in data-driven viticulture.