Pesticide-free agricultural strategies need new tools for disease prevention. Better than early detection of disease, detection of conditions favorable to their appearance can be a progress. In the case of fungal diseases, the presence of water on the plant surface is necessary. In order to detect remotely this presence early and at the scale of a crop field, we propose a low-cost solution based on laser reflection. Here, experimental results in a controlled environment are presented on both hydrophobic and hydrophilic leaves (rapeseed Brassica Napus and grapevine Vitis Vinifera, respectively). We first assess the water detection on a leaf surface by recreating the dew formation process. We next evaluate the influence of the scanning measurement and leaf inclination on the detection to get closer to in-field conditions. Results show that this method is very sensitive on both types of leaves. Water detection is possible from a low surface coverage with a high temporal precision at 1 m. In the hydrophobic case, water on a leaf surface leads to an increase of the detected signal up to three times compared to a dry leaf. The corresponding minimum surface coverage detectable at 1 m is evaluated at 1.6% thanks to 2D ray-tracing numerical simulations. In the hydrophilic case, on the contrary, water on a leaf surface leads to a decrease of the detected signal by almost half. For both types, the dew detection delay is contained under 5 min and can be improved. Finally, the presented results pave the way to a field application.