Detailed land cover datasets were developed for several ecologically-important areas in Clark County, Nevada, by classifying DigitalGlobe’s QuickBird satellite high-resolution (3 m2) multispectral imagery using Visual Learning Systems’ Feature Analyst feature extraction software. Over 1,000 vegetation field samples were collected at the stand level in the Red Rock Canyon National Conservation Area, and the Coyote Springs, Piute-Eldorado Valley, and Mormon Mesa Areas of Critical Environmental Concern. The field samples were classified to the National Vegetation Classification Standard, Version 2 hierarchy at the alliance level and above. Vegetation feature extraction models were developed on the basis of the spectral and spatial characteristics of selected field samples by using the Feature Analyst hierarchical learning process. Individual model results were merged to create one dataset for the Red Rock Canyon National Conservation Area and one for each of the Areas of Critical Environmental Concern. Field sample points and photographs were used to validate and update the dataset after model results were merged. Non-vegetation data layers, such as roads and disturbed areas, were delineated from the imagery and added to the final datasets. The resulting land cover datasets contain substantially more spatial and thematic detail than was previously available.