Conventional remote sensing classification techniques that model the data in each class with a multivariate Gaussian distribution are inefficient, as this assumption is generally not valid in practice. We present a novel, independent component analysis (ICA) based approach for unsupervised classification of hyperspectral imagery. ICA, employed for a mixture model, estimates the data density in each class and models class distributions with nonGaussian structure, formulating the ICA mixture model (ICAMM). We apply the ICAMM for unsupervised classification of a test image from the AVIRIS sensor. Four feature extraction techniques namely principal component analysis, segmented principal component analysis, orthogonal subspace projection and projection pursuit have been considered as preprocessing steps for reducing the data dimensionality. The results demonstrate that the ICAMM significantly outperforms the K-means algorithm for land cover classification of hyperspectral imagery implemented on reduced data sets. Moreover, datasets extracted using segmented principal component analysis produce the highest classification accuracy.