Accurate estimation of water consumption requires detailed information on vegetation types, including vegetable that is increasingly becoming one of the most important crops in China and many parts of the world. In current paper, a technique for rapid and accurate extraction of vegetable field (both greenhouse and open field) information from Landsat TM image is developed and tested. Through conducting field experiments and analyzing the Landsat TM images, we obtain the spectral characteristics of the film covered greenhouse vegetable and build a model to extract vegetable fields from the Landsat TM images. Applying this technique to Luancheng County of Hebei Province, China, we calculate the total area of the vegetable field being 208 ha. This number compares well with the result of our field survey. Based on the vegetable field area and the vegetable water consumption rate, we arrive at an estimation of the total vegetable water consumption in Luancheng County being 2.6*106 m3. The analyses also show that, in Luancheng County, the vegetable consumes about 4% of the total agricultural water use. The technique developed here provides an effective way for deriving vegetable field area from Landsat TM data and estimating the regional vegetable water consumption.