A key building block in the design of ultra-reliable communication systems is a wireless channel model that captures the statistics of rare events and significant fading dips. Ambitious reliability objectives, on the order of 10 5 − 10 9 packet error rate, only make sense when they are related to a statistical model of the environment in which the system is deployed. In this study, we propose a novel methodology based on extreme value theory (EVT) to statistically model the behavior of extreme events in a wireless channel for ultra-reliable communication. This methodology includes techniques for fitting the lower tail of the distribution to the generalized Pareto distribution, determination of optimum threshold over which the tail statistics are derived, derivation of optimum stopping condition on the sufficient number of samples required to apply EVT, and finally an assessment of the validity of the derived model. The analysis demonstrates that the proposed algorithm decreases the number of required samples for modeling the tail statistics by about 7 × 10 5 , and provides the best fit to the collected data and performs much better than the conventional methods based on the extrapolation of the average statistics in the ultra-reliable regime.