Investment in the stock market, despite its high return potential, is always associated with significant risks. The quality of a stock portfolio significantly depends on the future performance of the market and the selected industries for investment. This research aims to reduce this risk and improve returns by presenting an integrated four-stage approach for constructing an industry-based stock portfolio in the Iranian capital market.

In the first stage, using Dempster-Shafer theory and three technical depth indicators, MACD, RSI, and SMA, the future state of the market (bullish or bearish) is predicted. If a bearish trend is predicted, investment is redirected to low-risk assets such as Islamic Treasury Bills. If a bullish trend is predicted, the second stage begins, and industries are grouped using the K-Means clustering algorithm to create a basis for diversification. In the third stage, industries within each cluster are ranked using the TOPSIS method based on the same technical indicators. Finally, in the fourth stage, the top industry from each cluster is selected, and an equal-weighted stock portfolio is formed.

Model evaluation over a six-half-year period (from early 2021 to the end of 2022) on the top ten industries of the Iranian capital market showed that the model's accuracy in predicting market trends was 83%. Furthermore, the portfolio formed based on the model's selected industries had a significantly higher return than a portfolio consisting of the entire market (equal-weight index). The results of this study indicate the desirable performance of the proposed model, and it is suggested that active investors use this framework to form an optimal portfolio.