Hyperspectral airborne remote sensing provides time-efficient and high resolution data for middle-scale areas. Using the hyperspectral imaging sensor AISA Eagle attached to a motor-glider we observed the seasonal development of rocky intertidal macroalgal assemblages at the island of Helgoland (Germany, North Sea). The image analysis consists of radiometric and geometric correction, and image data reduction, implemented by Minimum Noise Fraction and Pixel Purity Index. Parallel field spectroscopy during flight campaigns is used for atmospheric correction. Field spectroscopy was used to build a spectral library for supervised classification. The classification results were compared with intensive ground mapping for verification. The major challenge of the classification is to distinguish different algal species or genera and mixtures of them which are abundantly present in natural assemblages based on their spectral reflectance. Our results from field spectroscopy of dominant brown algae of the order Fucales and Laminariales indicate that species of the same genus are indistinguishable while differentiation between brown algal genera (Fucus, Sargassum, Laminaria) is possible. We present an approach that uses a Gaussian Mixture Model as probabilistic classifier based on the spectral peak ratios of selected wavelengths.