Dominant rhythm in electroencephalographic (EEG) records is seen at the posterior to occipital region as a primary component of background activity during waking state with closed eyes and is decreased or disappeared by the exogenous factors such as visual stimuli to eyes and so on. Visual evoked potential (VEP) can also be seen in EEG at the posterior to occipital regions during photic stimulation (PS). Frequency components of VEP are depended upon the frequency of stimuli and that of dominant rhythm in case of healthy adult is around 10 Hz. Therefore, components of VEP and dominant rhythm are almost overlapped when the frequency of photo stimuli is near around 10 Hz. VEP component can be extracted from the background activity by using the averaging method, but the accurate estimation of dominant rhythm component in such condition has difficulties due to the overlapping of both components in frequency domain. Some of the authors have proposed the EEG model with Markov process amplitude (MPA EEG model) in the past. The MPA EEG model has possibilities to separate the components that construct the original EEG into each one in the frequency domain. In this study, component decomposition of VEP and dominant rhythm of recorded EEG was done by using the MPA EEG model. EEGs with PS were recorded from three healthy young adults. Five seconds continuous EEG time series with PS and without PS were selected from the original data, and were transferred to the periodogram information by the FFT method. Then, the model parameters were calculated. The initial values of model parameters were determined from the periodogram of raw EEG during PS, and were optimized by Fletcher-Powell method. In the original data under PS with 10 Hz, VEP component and dominant rhythm component were overlapped each other. Proposed method decomposed the original data into five components; first harmonic VEP, second harmonic VEP, dominant rhythm, slower noise and others. Characteristics for the depression of dominant rhythm and the amplitude of VEP were quantitatively analyzed from the decomposed component by the MPA EEG model. Effectiveness of the proposed decomposition method was also investigated.