In wireless optical communication systems, the transmission of optical signals via the channel (air or water) is affected by absorption and scattering. These reduce the signal strength (attenuation) and transmission distance of the signals. In pure water, the blue-green region of the visible light gives low attenuation. Some models have been developed to characterize the underwater optical channel such as Beer Lambert’s law, Radiative Transfer equation and Monte Carlo model. In underwater optical communication, optical power meters are an invaluable tool in the determination of attenuation coefficients. However, optical power meters for underwater optical communication are very expensive. There is a need to be able to determine the attenuation in the underwater optical communication channel at a low cost, especially in the absence of underwater optical power meters. In this paper, we present an alternative low-cost experimental method of obtaining the approximate attenuation coefficient of the visible light beam in underwater optical wireless communication without the use of optical power meters. A wireless visible light communication system was set up experimentally for underwater measurements involving an oscilloscope as the only measuring device. The system uses sub-carrier frequency modulation; a white light-emitting diode array for the transmitter, and a solar panel at the receiver front end. A theoretical transmission model was developed from the experimental setup based on the line of sight method in an unbounded medium, Beer Lambert’s law, and the received sub-carrier signals; in order to provide an alternative method of determining approximately the attenuation coefficient of the underwater medium. Experiments were performed in air, clear water and in saline water. The attenuation in the air was used as a reference upon which attenuation in the clear water and the saline water was based. The saline water has a salt concentration of 6.7 g/100 mL by weight and a total dissolved solid of 86.2 ppt. The trend of the measured received sub-carrier signals showed deviation from the developed theoretical model, and the model was therefore adjusted to conform to the experimental results. From the adjusted model, the attenuation coefficient of 0.0007379 cm-1 and 0.02447 cm-1 were obtained for clear water and the saline water respectively. The method is simple, straightforward, easy to set up in a laboratory, low cost and can be applied to visible light of any wavelength.