Anomaly detection plays a major role in hyper-spectral remote sensing target detection algorithms. The power of anomaly detection in these algorithms is its independence from prior knowledge about the target’s spectrum and the insensitivity to atmospheric corrections on the hyper-spectral image. This article describes a simulation for anomaly targets in hyper-spectral images. The simulation is based on mathematical concepts of statistical anomaly detection algorithms that model the background and discriminate anomalies from background pixels in the hyper-spectral images. With this simulation, anomaly detection algorithms can be tested and redeveloped to cope with anomaly targets of different strengths in order to improve their performance.