This study estimated leaf area index (LAI) of individual urban trees as a function of spectral features derived from airborne hyperspectral data. Candidate features included spectral indexes, principal components, and calibrated reflectance values. Hyperspectral images were acquired over Provo, Utah area, and LAI of 204 deciduous trees was measured in the field. These tree canopies were identified on the images, and spectral features were extracted using both whole canopy and mean-lit spectra techniques. Multiple regression and artificial neural networks were used to model leaf area and determine which spectral features were most strongly related to it. Results established that simple hyperspectral vegetation indexes explained more variation in urban tree LAI than either principal component scores or simple band reflectance values. The neural network model trained with a subset of those indexes explained more variation in LAI (R2 = 64.8 percent) than any of the multiple regression models tested.