Structural testing criteria help the tester in the generation and evaluation of a test case set T.They are predicates to be satisfied to consider the testing activity ended and generally require the execution of a set P of paths,capable of exercising certain elements in the program under testing.Determining P is an important and hard task,and its automation is strongly desirable fo easing the criteria application.This task can influence on the efficacy and on the testing effort and costs.This work explores the use of diverse programs characteristics to propose strategies for selection of testing paths.The work also describes a module that implements a framework for representation and automation of those strategies.Using this module, a testing procedure is presented and a strategy, that uses the number of predicates to select paths, is evaluated. The obtained results give some information about the main advantage of this strategy:to easy the automatic test data generation by reducing the number of selected infeasible paths.