Constant False Alarm Rate (CFAR) detection is an important concept in radar detection problems. CFAR processors are useful for detecting radar targets in background for which some parameters of the statistical distribution are not known and may be non stationary. The conventional method for producing CFAR property in radar is to set the detection threshold adaptively and equal to the scaled version of the noise power estimated by processing the resolution cells surrounding the cell under test to achieve the desired constant false alarm probability. Various algorithms have been proposed to keep the false alarm rate constant in different environmental conditions, such as the presence of interfering targets or clutter edge. These algorithms can be divided into single pulse CFAR detectors (one dimensional CFARs) and multiple pulse CFAR detectors (two dimensional CFARs). Multiple pulse CFAR detectors are detectors that use the samples of multiple pulses in each range cell for target detection. A conventional processing method in these detectors is to integrate the pulses incoherently and apply the one dimensional CFAR detector to the result. Although CFAR detection has been studied widely for monostatic radars, reviewing the publications in this field shows that it has not been considered seriously for MIMO radars. MIMO radar is defined broadly as a radar employing multiple transmitters and receivers and capable of jointly processing signals received at multiple receive antennas. MIMO radar may be configured with its antennas collocated or widely separated over an area. In this thesis, CFAR detection for MIMO radars with widely separated antennas in Gaussian noise is considered. After modeling the received data in a single pulse MIMO radar with widely separated antennas, some of the one dimensional CFAR detectors from Averaging and Ordered statistics families have been generalized in several ways for MIMO radars. Some of these methods are similar to those employed to generalize one dimensional CFAR detectors for processing multiple pulses in monostatic radars. Parameters of these generalized detectors have been found using mathematical analysis. Also, performance of these detectors has been evaluated by means of detection and false alarm probability calculation in homogeneous background and computer simulation results in nonhomogeneous background. Comparison of different detectors’ performances shows that two dimensional CFAR detectors used for processing multiple pulses in monostatic radars are not perfect for MIMO radars. Instead, the algorithms that estimate the total noise power by averaging the noise power estimations in MIMO radar’s different receivers have a better performance in most . In addition, it has been shown that proposed detectors which are Keywords Radar, MIMO radar, Detection, CFAR detection