Software plays an important role in modern society. New software features have to be completed and tested rapidly due to business and user needs and goals. To ensure the quality of modified software components, new test cases must be generated and included in the current test pool. Furthermore, test pool size often grows too large to test all cases. A considerably large test pool typically consumes an exorbitant quantity of time, leading to inefficient regression testing. Test suite reduction is often used to solve this problem by removing redundant test cases. Thus, time and resource constraints make it necessary to minimize the test suite. That is, test pool size should be efficiently reduced, while the remaining test cases should be adequate to provide the same coverage as the pool. The main purpose of this paper is to propose four weighted combinatorial algorithms and a three‐objective binary integer linear programming (ILP) model to improve fault detection effectiveness. We first consider both the control flow and data flow as our testing criteria and combine them by using a weighted means with four existing test suite reduction algorithms. A genetic algorithm is used to find the best combination of weighting factors to assign to the two testing criteria. Experiments based on real subject programs are presented. Experimental results demonstrate that 77.5% of our proposed weighted combinatorial algorithms have an even or better performance in suite size reduction, 83.75% of our approaches perform even or better in fault detection effectiveness, and 82.5% of our approaches have even or better effectiveness in fault‐to‐test ratio. For the three‐objective binary ILP model, test suite size is significantly reduced without any accompanying loss in fault detection ability.