Adaptive beamforming increasingly finds application in new communication and remote sensing systems. In the beamforming, a weight vector is applied to the signals received by an array of antennas. The weight vector is usually computed to maximize the desired received signal and suppress interference and noise sources. However, the location of the desired signal source may vary or may only be known within a certain region. In such case it is beneficial if the beamforming can incorporate this uncertainty. Here, a new robust beamforming (RB) algorithm is presented and demonstrated which incorporates the location uncertainty through a probability density function for the location of the desired signal. The approach is demonstrated using an array of 10 patches at 2.4 GHz and four cases of the probability density function.