Applying high-reliability and low-latency scenarios of the fifth-generation mobile communication technology (5G) to large-capacity and low-latency services in the power system requires meeting the communication reliability requirements specified by the power system. However, reliability metrics based on packet loss rate fail to capture the time-dependent characteristic of time-varying wireless channels and the healthy transmission duration required for different ultra-reliable and low-latency communication (URLLC) services. Therefore, to analyze the performance of channels within a given time interval, this study employs survival analysis methods to combine URLLC key technologies with failure rate functions in reliability theory. By utilizing the received signal strength (RSS), theoretical distribution models and data-driven models are established to predict the channel reliability within the next subframe. A case study is conducted by using of the UMa-NLOS-CDL channel model to compare and analyze the proposed methods. Furthermore, considering different rain fading conditions, the reliability of multiple-input multiple-output (MIMO) and multiple-input single-output (MISO) channel systems is compared and analyzed