This paper focuses on an efficient approach for designing multi-static arrays for millimeter-wave imaging, based on the k -space or Fourier-spatial domain characteristic of imaging systems. Our goal is to decrease the redundancy of the data measured by each antenna and to improve the resolution of the reconstructed image. The proposed technique is based on determining the role of each transmitter and receiver, in collecting the data from each voxel of the target in k -space domain and then rotating the transmitters' beams to measure the desirable information. The effect of non-uniform redundant k -space domain frequency samples that act as an undesirable filter is compensated using a modified SAR back-projection algorithm. The experimental and simulation results are presented and compared with that of a sparse multi-static array with the same number of transmitters and receivers. Our simulations and measurements show significant improvement in terms of overall quality and edge preservation in the reconstructed images. Also, the obtained results demonstrate that using the proposed structure and algorithm, the average improvement in peak-signal-to-noise ratio, structural similarity index measure, and digital image correlation metrics of 3.03, 0.22, and 0.173 dB, respectively, are acheived.