The purpose of this study is to propose an equating method for equating two latent ability scales (scale θ* & θ) by using subjects' estimated ability scale values and test information functions. This method uses asymptotic property of maximum likelihood estimator's distribution. The practical application of this method is demonstrated by using a part of the test for word meaning comprehension (Shiba & Noguchi, 1982). As a result the computer simulation shows that this method has smaller error of estimate than Noguchi's (1983).