The south-west monsoon (June–September) is the major rainy season over India. Information about the occurrence of precipitation and the expected quantity at a specific place is important in many sectors of human activity. In this study, objective methods are developed to forecast the probability of precipitation (POP) and provide the quantity of precipitation forecast (QPF) over Delhi. As the onset of the monsoon at Delhi is around 30 June, the models are developed for the months of July, August and September (JAS) using surface and upper-air data for the period 1985–90 and tested with data from JAS for 1994 and 1995. A multiple linear regression equation is developed to forecast the POP and multiple discriminant analysis is used to produce the QPF in terms of one of four groups (0.1–1.0; 1.1–10.0; 10.1–30.0; and ≥ 30.1 mm). The QPF model is used only if precipitation is expected to occur (the POP forecast is turned into a categorical forecast). The categorical forecasts based on the POP exhibit positive skill scores consistently with both the development and independent data sets. The model for QPF also performed satisfactorily