A new band selection method for hyperspectral images, namely Maximum Band Screening (MBS), is proposed. Considering a band image as a band vector, a spectral similarity measure is used by MBS to measure the dissimilarity between a band and the bands selected, and the most distinctive and informative bands are selected form the original bands. In order to demonstrate its utility, an anomaly detection algorithm is developed, which first extracts the anomalous target spectrum from the image, followed by noise whitening, then implements MBS to select bands that participate in the subsequent Adaptive Cosine Estimator (ACE) target detection. As shown in the experimental result on the AVIRIS dataset,only 8% of the full bands are selected by MBS, with which the detection performance of ACE has been improved greatly over that by other two band selection methods.