Abstarct: Automatic age range estimation is one of the artificial intelligence issues that has been studied for many years due to its importance. Interesting studies also have been done to distinguish the age range. But despite many studies, the accuracy of these systems is still a challenge for speaker age estimation systems. Therefore, it needs more studies and new works in this field, especially for Persian speakers. The goal of these systems is to increase the accuracy of age estimation. Therefore, we are always facing two main challenges in this area which are as follows: finding a suitable method for extracting feature from the speech signal and finding a reliable method for data classification. The main purpose of this study is to estimate the age from speech signal using the Power Normalized Cepstral Coefficients (PNCC) feature for the feature extraction part in an estimation system And finding the best amount to have a lower error rate. Also, the performance comparison of this feature with other common features such as MFCC, MHEC, and BFCC has been examined. SVM classifier is also used for classification. This method has been performed in three, four, five, and six classes, The best result has been obtained in the classification of three classes using the PNCC feature extraction method, which has an accuracy of 78.6%.