An equivalence is established between generalized master equations and continuous-time random walks by means of an explicit relationship between(t), which is the pausing time distribution in the theory of continuous-time random walks, and(t), which represents the memory in the kernel of a generalized master equation. The result of Bedeaux, Lakatos-Lindenburg, and Shuler concerning the equivalence of the Markovian master equation and a continuous-time random walk with an exponential distribution for(t) is recovered immediately. Some explicit examples of(t) and(t) are also presented, including one which leads to the equation of telegraphy.