The main interest of this study is to find a system that could detect typical signs of drowsiness progression and warn a car driver before driving behavior becomes dangerous. An early detection of impaired conditions due to drowsiness would probably lead to a reduction in traffic accidents. In this matter, a lot of researches have already been done but although many detection devices are available on the market today, the validity of most of them needs to be confirmed. The aim of this project is to develop and test a model for detection and categorization of driver drowsiness by evaluating EOG data from a number of test subjects. The data were recorded using an advanced module system and used to simulate normal and sleepy drivers. The empirical mode decomposition method is proposed as a signal decomposition tool. This kind of methods is useful for the analysis of natural and non-stationary processes. Some parameters are calculated for each intrinsic mode function (IMF). EMD is proved to be adaptive and highly efficient in the analysis of such signals and the proposed parameters provided significant differences between normal and sleepy status.