This study tested the long-term dependence and mean reversion of monthly stock prices in Korea using test statistics that reflect the characteristics of time-series autocorrelation. First, applying the R/S statistic, which allows for short-term time-series dependence arising from microstructural imperfections in the capital market, the results showed that monthly stock prices in Korea do not follow a random walk model and contain a component that persists over the long term. Furthermore, applying an autoregressive model of monthly returns, mean reversion of past stock prices was observed. These test results carry the financial implication that historically realized stock returns can serve as useful information for predicting future returns.