In this paper, classification of brain wave using real-world data from Parkinson�s patients in producing an emotional model is presented. Electroencephalograph (EEG) signal is recorded on eleven Parkinson�s patients. This paper aims to find the �best� classification for brain wave patterns in patients with Parkinson�s disease. This work performed is based on the four phases, which are first phase is raw data and after data processing using statistical features such as mean and standard deviation. The second phase is the sum of hertz, the third is the sum of hertz divided by the number of hertz, and last is the sum of hertz divided by total hertz. We are using five attributes that are patients, class, domain, location, and hertz. The data were classified using WEKA. The results showed that BayesNet gave a consistent result for all the phases from multilayer perceptron and K-Means. However, K-Mean gave the highest result in the first phase. Our results are based on a real-world data from Parkinson�s patients. � Springer International Publishing Switzerland 2015.