The software development life cycle includes software testing as a critical step. Oracle testing is necessary to determine if a particular test case detects an error because the purpose of testing is to ensure that the application adheres to its specifications. This test leads to the consumption of time and cost and may lead to errors in which the tester falls. This research aims to present a proposed system based on the automation of the Oracle test using intelligent techniques. The model was trained using the random forest algorithm and using the convolutional neural network. This proposed system is considered one of the box techniques. Test blacks that depend on the output of the system being tested.