In order to properly rank the priority of test cases from the requirement's perspective, this paper introduces three impact factors: Concerned-Requirements coverage, test case importance degree and test case failure rate. Meanwhile, three weight factors α,β and γ are introduced to balance the three impact factors. This paper designs on-line estimating methods and algorithms based on the concerned-requirements coverage and the test case failure rate. Based on those metrics, a multi-objective optimization based test case prioritization on-line adjustment strategy is developed. The strategy is able to adjust the priorities of test cases dynamically using the feedback information collected in the test process, and thus can meet the coverage criteria earlier and cover test cases of importance and high failure rate. This strategy can also resolve the multi-object test case prioritization problem by detecting more severe bugs earlier. Experimental results show that, compared with random test, the traditional single-object test and the test with deterministic test case prioritization, the presented strategy can complete the test with equal quality by shorter time, thus improves the testing efficiency.