It is expected that the development of test suites from models can contribute substantially to reducing the human effort and to increase the effectiveness of the generated tests. Means for tests’ effectiveness (ideal) the percentage of existing defects found by these tests. Most of these techniques is based on state machines and mostly directed to testing the functionality. However, there is a need to be able to test highly interactive systems, such as smartphones and tablets, from a description of its human computer interface. The goal of the dissertation is to make a first evaluation of a technique aimed to generate test suites for test of human computer graphic interface. For this purpose was developed and evaluated its effectiveness, a tool called Easy, using use cases tabular and state machine for the automatic generation of the suite tests. The use cases are described in natural language restricted. From this description, the tool builds a state machine, and then uses this to generate scenarios. By construction scenarios will be in accordance with the use cases. Each scenario corresponds to a test case. The scenarios are presented to the user in natural language, allowing the visualization of them before the generation of the final scripts tests. The generated scripts are intended to a running automated tool geared to testing graphical interfaces. In this work, we used the UI Automation tool, responsible for running tests on applications for the iOS, operational system for iPhones, iPads and iPod touchs. The effectiveness of the procedure was evaluated in a real application, available in the online store applications App Store. In addition, HCI tests were performed in order to evaluate the influence on the cost of production of the test suite.