Multiple linear regression models were developed for prediction of pearl millet dry biomass and yield under agroclimatic conditions of Delhi. A stepwise regression analysis was carried out to select the best agrometeorological (absorbed photosynthetically active radiation, evapotranspiration and thermal interaction rate) and spectral indices (infra red/red ratio aod normalised difference) which explained maximum variation in dry biomass and yield. The R(2) values obtained were 0.973 and 0.995 for dry biomass and yield models, respectively.