In this work, we apply single neuron method to relieve freeway traffic congestion. We consider a freeway composed of cells and entry/exit ramps, and formulate the ramp metering problem as a density tracking process. The cell transmission model (CTM) is firstly formulated and ramp control objective is determined. Based on CTM and single neuron, a freeway ramp metering system is then designed, and the learning algorithm of single neuron is given in detail. Finally, the ramp metering system is simulated in MATLAB software. The results show that this method can effectively deal with this class of control problem, and can achieve a perfect density tracking performance. This ramp metering can eliminate traffic congestion and maintain traffic flow stability.