In this letter, we propose a layered video streaming technique that combines network coding (NC), multiple-input multiple-output (MIMO), and hierarchical modulation (HM) over cellular relay networks. We provide a performance analysis of different transmission modes of NC and MIMO in terms of the error rate in video layers, which is reflected in the total layered-video quality. The HM is used to differentiate the error rates among video layers. The simulation results show that the NC, the MIMO spatial multiplexing (SM), and the combination of both the NC and MIMO-SM gives video-quality gains of about 1.9dB, 6dB, and 12dB maximally, respectively, compared to the conventional single-input and single-output (SISO) single relay network (SRN) system.