The use of airborne hyperspectral sensors for urban analysis represents a significant advance in remote sensing. The greatest challenges to effectively using urban hyperspectral imagery include (i) managing the natural spectral complexity of urban fabrics, (ii) selecting optimal feature sets from an enormous number of candidate features, (iii) designing classifiers that are minimally affected by the curse of dimensionality, and (iv) reducing the computational burden of hyperspectral algorithms on large images. To address these challenges, there have been several promising methodological advances in urban hyperspectral analysis. Examples include multiclassifier systems, decision fusion processes, support vector machine algorithms, object oriented approaches to classification, and discrimination via morphological profiles. Unfortunately, few of these advances are leveraged in commercial software packages, thus limiting their practical utility to the practitioner.