It would be a great advantage if remote classification of species is possible under a variety of vegetation conditions, such as various levels of moisture stress. The primary goal of this study is to investigate the use of hyperspectral data for the detection of soybean (Glycine max) from weeds, specifically sicklepod (Senna obtusifolia) and cocklebur (Xanthium strumarium) at various levels of moisture stress and to determine the effects of moisture stress on automated weed detection systems. A secondary goal of this study is to investigate the use of hyperspectral data for the detection of moisture stress within a given vegetative species. Two feature extraction techniques were investigated, including the use of spectral bands' amplitudes and the use of discrete wavelet transform coefficients. These features were used in a traditional maximum-likelihood classification system. Experimental results showed that higher moisture stress (less moisture) corresponded to higher weed detection accuracies.