5G cellular networks are of high user demand in mobile data services. A millimeter wave technology is considered a potential one for 5G mobile networks, as it allocates more bandwidth. The next generation of wireless networks i.e Millimeter wave offers multi-gigabit data rates for many mobile users with higher attenuations. The major drawback in high-frequency communication systems (mmWaves) is Phase noise. These Phase noises are due to non-linearity conditions of local oscillators, which introduce Common phase error (CPE) & Inter-carrier interference (ICI). To overcome the phase noise effect in mmWave systems, the Multiple-Input & Multiple-Output (MIMO) beamforming is used for higher order data transmissions. In the situation of the common oscillator, the phase noise at MIMO-OFDM has the same influence on the transmitter and receiver sides in terms of Error vector magnitude. But in the independent oscillator case, it has different levels of influence. The measurement of channel metrics such as Average Power Delay Profile, Power Elevation Profile, and Power Azimuth Profile are employed to reduce the Phase noise in millimeter MIMO-OFDMA systems. A Unitary Space Alternating Generalized Expectation-maximization (U-SAGE) algorithm is proposed for channel measurements based on Phase noise reduction in millimeter waves. This paper discusses the reduction of Phase noise greatly in independent oscillator cases using U-SAGE Algorithm.