In this paper, we discuss the issues in automatic recognition of vowels in Persian language. The present work focuses on new statistical method of recognition of vowels as a basic unit of syllables. First we describe a vowel detection system then briefly discuss how the detected vowels can feed to recognition unit. According to pattern recognition, Support Vector Machines (SVM) as a discriminative classifier and Gaussian mixture model (GMM) as a generative model classifier are two most popular techniques. Current state-ofthe- art systems try to combine them together for achieving more power of classification and improving the performance of the recognition systems. The main idea of the study is to combine probabilistic SVM and traditional GMM pattern classification with some characteristic of speech like band-pass energy to achieve better classification rate. This idea has been analytically formulated and tested on a FarsDat based vowel recognition system. The results show inconceivable increases in recognition accuracy. The tests have been carried out by various proposed vowel recognition algorithms and the results have been compared.