The mechanical strength of solid catalysts is distributed over a wide range of values. The suitability of the normal, lognormal and Weibull distributions to model the catalyst strength variation was judged by three goodness-of-fit criteria: the coefficient of determination, Akaike information criterion and Kolmogorov-Smirnov test. It is concluded that the Weibull distribution is most universal to represent the catalyst strength data, though sometimes it may not be the optimal candidate. It is elucidated that the catalyst strength variation results from the brittle fracture nature of the mechanical failure of solid catalyst materials. The low-strength/probability part of the catalyst strength distribution is the key domain for the mechanical reliability, while the mean strength that has traditionally been taken as a measure of the catalyst strength is of less importance. It is also pointed out that a mechanical strength distributed in a narrow window, i.e., with a high Weibull modulus, is beneficial to industrial applications of solid catalysts.