While much of the previous research on mutual funds has concentrated on finding the relationship between the investment style, the past performance and the future performance of funds, very few of the studies have paid attention to the effect of a mutual fund manager’s execution of investment style on fund returns. Using return-based analysis methodologies for measuring the style consistency of Chinese mutual funds, this thesis demonstrates that the less style-consistent funds tend to produce higher future risk-adjusted returns than more consistent mutual funds, even after controlling for past performance and net asset value (NAV). Further, these findings are robust across mutual fund investment style classifications, test period intervals (one-year or one-quarter interval), and the model used to calculate the expected returns (four-factor model and Sharpe’s style analysis model). This thesis also documents the performance-persistency effects that exist in Chinese mutual funds, which remain persistent even under the condition of style consistency. More importantly, the research discovered that at a time of change in the Chinese stock market, the negative correlation between style consistency and future performance becomes weaker. The study concludes that style consistency does matter for mutual funds’ future risk-adjusted returns and that there is a significant negative correlation with mutual funds’ future risk-adjusted performance in the longer term (i.e., over the entire test period). Moreover, this connection is distinct from those related to the past risk-adjusted performance and NAV of mutual funds. It is also clear that a significant negative correlation between style consistency and the future risk-adjusted return does exist in Chinese stock and asset allocation mutual funds, even after adjusting for the investment style of the fund. Finally, this thesis provide a mutual funds picking strategy for investors base on the main findings of this study, which can provide significant positive alpha at each year during the test period.