Data envelopment analysis (DEA) has been abundantly applied for portfolio selection, albeit in a single-period framework. However, in reality, portfolio selection and management is a multiperiod process stretched over long investment horizons. Investors often reallocate their capital and rebalance the portfolio based on the changing market conditions and performance of the assets and portfolio as a whole. Since the performance of the assets constituting the multiperiod portfolio changes from one period to another; therefore, it becomes imperative to employ DEA in a multiperiod framework for an efficient portfolio selection. To this end, this article proposes a two-stage multiperiod efficient portfolio selection (MPEPS) approach wherein the assets’ return rates are characterized as trapezoidal fuzzy numbers. The fuzziness in the assets’ returns is handled using the credibility theory. In stage-I, the relational network DEA is employed to evaluate the assets’ periodwise and overall efficiency with variance, entropy as the inputs and return, liquidity as the outputs. In stage-II, an MPEPS model is proposed to maximize the terminal wealth and efficiency, and minimize the CVaR of the portfolio subject to the capital budget, lower and upper bounds on the capital allocated to an asset, no short selling, cardinality, and contingent constraints. The normalized weighted sum approach is employed for solving the proposed model. Moreover, a comprehensive numerical illustration is presented to demonstrate and validate the proposed approach through implementing several investment schemes representing different investor attitudes.