An automated, self-adaptive framework prioritizes software testing in a consistent and effective manner. A metric evaluates past test execution information for assigning regression testing priority. The metric may be calculated with reference to one or more of the following factors taken in combination: requirement, coverage, history, and cost. The requirement factor considers customer-assigned priority of testing the code, complexity of implementing the code, and proneness of the code to faults. The coverage factor considers code coverage, feature coverage, and common usage rate. The history factor considers previous bug found rate, case stable rate, and priority to calculate. The cost factor considers test case execution time, and step length. A value of each factor for one test case is measured according to that test case and is not related to other test cases. The calculation result representing the metric for each test case determines a priority of the test case.