This paper introduces a method to improve the azimuth angle resolution using MIMO-FMCW radar. When using a MIMO-FMCW radar, a 2D radar image composed of a range axis and an azimuth axis can be obtained. The range resolution is determined by the bandwidth, and the azimuth resolution is determined by the length of the virtual antenna array and the number of virtual antenna elements. To improve the azimuth angle resolution while avoiding aliasing, in this paper, the virtual antenna was placed wider with non-uniform spacing. Then, deep learning technique was applied to reduce the side lobe effect. The proposed method was verified through experiments using simulation signals and emulation signals based on measurements.