The aim of this thesis was to conduct an experiment in which it was sought to determine whether the moisture content of live vegetation can be determined from the spectral characteristics of plant leaves. In order to determine the spectral characteristic, the AMS AS7265x multispectral sensor was used, which is adapted for spectral analysis in the visible and near infrared spectrum. The sensor has a spectral response defined in the wavelength range from 410 nm to 940 nm divided into eighteen spectral channels allowing us to perform detailed spectral analysis. For the actual value of the moisture content of the leaf, the value read from the KERN DBS halogen moisture analyzer was used. The conducted experiment lasted for eight days. Three plants with different moisture levels were included in the experiment - a plant watered twice a day, a plant watered six times a day, and a plant that was not watered since the beginning of the measurement. Data on soil moisture, leaf moisture, and transmission coefficients on each of the eighteen spectral channels that the used optical sensor can measure were collected each day. After data collection, the functional dependence of leaf moisture content and transmission coefficient for each of the eighteen channels was analyzed, and models were made from the obtained data that predict the water content in the leaf from its spectral characteristics using supervised machine learning algorithms. The best model had a coefficient of determination of 0,427. The resulting model can be used for rapid and non-invasive assessment of live vegetation moisture content and can find its application in agriculture as well as fire prevention.