An application of the neural networks model is considered to estimate Research Octane Number of gasoline. It is one of the most important qualities of gasoline, and shows the anti-knocking performance of the fuel. Three methods for estimating Research Octane Number are developed in this study; the empirical method using Research Octane Number of each component of gasoline, the statistical analysis method, and the neural networks model. It is shown that the neural networks model is effective in estimating Research Octane Number.