Outlined is a method for a short-range probability forecast of the minimal and maximal air temperature, steady rains, and showers in Moscow with a lead time of 12, 24, and 36 hr. The probabilities of temperature and precipitation are estimated based on discriminant functions. Predictors in the discriminant functions are obtained from results of hydrodynamic forests of the fields of meteorological elements. Results of verification of methodological probability forecasts of temperature and precipitation show that they markedly exceed inertial and climatic forecasts.