QuickBird multi-spectral satellite imagery was evaluated for detecting weeds in waterways and wetlands in Texas. Plant species studied included giant reed (Arundo donax L.), waterhyacinth [Eichhornia crassipes (Mart.) Solms], and giant salvinia (Salvinia molesta Mitchell). Accuracy assessment results performed on the unsupervised classified maps for each species were: giant reed, 94.4% producer's accuracy and 100% user's accuracy; waterhyacinth, 100% producer's accuracy and 80% user's accuracy; and giant salvinia, 93.9% producer's accuracy and 92% user's accuracy. These results indicate that QuickBird imagery can be used successfully to map infestations of these three weeds.