Alzheimer's disease is a neurological disorder that is a major cause of dementia in the world. Early detection of Alzheimer's is very important to delay the degradation of brain function which can happen in a short time. Also, it is important to determine the right treatment for patients. Medical imaging such as MRI and CT-Scan can use for analysis of this deterioration. However, it requires a high cost for setting this protocol. As an alternative tool that can be used for analysis is an electroencephalogram (EEG). Therefore, in this study, we propose an early detection method for Alzheimer's based on EEG signals. To achieve this purpose, we used an open dataset consisting of 11 MCI patients and 16 normal subjects. Theta wave energy was analyzed and becomes a vector feature on each EEG electrode. Support vector machines (SVM) with 5-fold cross-validation are employed to evaluate the proposed method. Test results show the highest accuracy is 81.5%. This study shows that EEG analysis has the potential to be developed as supporting tools in the diagnosis of cognitive impairment