This paper addresses a dynamic portfolio investment problem. It discusses how we can dynamically choose candidate assets, achieve the possible maximum revenue and reduce the risk to the minimum level. The paper generalizes Markowitzs portfolio selection theory and Sharpes rule for investment decision. An analytical solution is presented to show how an institutional or individual investor can combine Markowitzs portfolio selection theory, generalized Sharpes rule and Value-at-Risk(VaR) to find candidate assets and optimal level of position sizes for investment (dis-investment). The result shows that the generalized Markowitzs portfolio selection theory and generalized Sharpes rule improve decision making for investment.