Software testing is an important activity for developing a quality end product. But it is expensive as it requires both time and cost to perform it during development phase and after its delivery. Therefore, it is required to develop some techniques in this direction in order to make software testing a desirable as well as an affordable activity. Hence, in this study we have proposed an effective method that minimizes the size and redundancy from an automated random generated test suite using data clustering technique. The proposed technique compared to other existing techniques achieves a substantial amount of reduction in the size of the test suite without affecting the fault detection capability of the test suite.