Exploiting the classical R/S and modified R/S analysis, we first reveal theevidence of long-term memory in liquidity, volume, and volatility. Thereafter,we estimate the fractionally integrated autoregressive movingaverage ARFIMA models by both the exact-maximum likelihood (EML) and themodified-profile likelihood (MPL) methods. Furthermore, based on the theoryof financial economics, we extend the simple ARFIMA models to the Multi-FactorARFIMA models by incorporating the mutual relationships among financial marketvariables and present the effectiveness of the Multi-Factor ARFIMA models infinancial markets.