The Receiver operating characteristic for a particular detection system is based on practical models for signals and noise using assumptions of underlying Gaussian probability density functions. The decision algorithm will be based on the assumption that the receiver buffer data have been processed to help the signal stand out from the noise such as filtering or averaging, and so on. Time-synchronous averaging is quite effective at improving the Signal-level measurement in periodic signals such as repetitive pulses in the sonar or radar. The chapter aims to develop a straightforward decision algorithm to detect Recursive Least-Squares signals in noise with a Constant false alarm rate (CFAR). CFAR detection is very desirable for situations where the background noise varies over time and the detection threshold must be maintained so that the false alarm performance remains at its designed operating point. The tracking filter, with its states and covariances available, is an ideal algorithm for linking to CFAR detection.