For the blindness of man made choice of the parameter of support vector machine , a ant colony optimization was applied to select parameters of SVM ; and a novel algorithm, ant colony optimization support vector machine was put forward. According to the measured data of the vibration signal of the engine valve, the engine valve faults diagnosis model based on the ACO-SVM was established, and was compared with the models based on GA-SVM and BPNN. Results show that the engine valve faults diagnosis model based on the ACO-SVM outperforms the models based on other two algorithms in learning efficiency and diagnosis accuracy, and is available for the faults diagnosis of engine .