In this paper, the thermal conductivity (k_nf) of cerium oxide/ethylene glycol nanofluid is extracted for different temperatures (T = 25, 30, 35, 40, 45, and 50 °C) and the volume fraction of nanoparticles ($\phi =$ 0, 0.25, 0.5, 0.75, 1, 1.5, 2 and 2.5%) and then k_nf is predicted by two methods including Artificial Neural Network (ANN) and fitting method. For both methods, the results have been presented and compared. The experiments showed that with increasing $\phi$ and temperature, the thermal conductivity ratio (TCR) of nanofluid increases. It was also observed that when the experiments are performed at high temperatures, the rate of increase in k_nf is much higher than the change in the same amount of $\phi$ change at low temperatures. An ANN with 7 neurons has a correlation coefficient very close to 1 and this proves that the outputs are compatible with experimental results. Also, it can be seen that the ANN could predict the thermal behavior of cerium oxide/ethylene glycol nanofluid more accurately.