Systems and methods for predicting crop diseases in a crop field using machine learning are disclosed. The method includes receiving one or more ambient weather parameters related to environmental conditions around the crop field from an environmental sensor; determining a disease risk indicator associated with the disease and the crop field based on the one or more ambient weather parameters, using a disease risk machine learning module trained on historical weather data from a weather database; receiving a pathogen incidence indicator related to the presence of a pathogen around the crop field from a pathogen sensor, wherein the pathogen is associated with the disease; and generating a disease score based on the disease risk indicator and the pathogen incidence indicator, where the disease score represents the severity of the disease around the crop field.