The purpose of this study is to model the factor stock selection investment strategy. Five indicators were selected for discussion in the financial reports and three investment models were established: single-factor model, equal-weighted two-factor model, and different-weighted two-factor model. In the single-factor portfolio, each factor is independently backtested and compared, and PB is the best factor indicator. The sample pool is filtered by market value and re-tested to get better annualized return and risk value. In the two-factor equal weighting model, the study combines two of the five factor indicators into a total of 10 combinations and optimizes for the investment cycle by switching to one-day trading after the release of the quarterly report and eliminating the worse-performing factor, EPS. the optimal combination for the optimized single-factor model is ROE, and the best combination for the optimized two-factor model is ROE + PB. In the two-factor model with different weights, we use the bisection method to find the best weighting, and after backtesting, the best annualized return combination is ROA+PB with a weighting of 2:3, and the annualized return is as high as 29.70%, which is the best stock picking model for this research.