Allometric model, in logarithm form, is generally fitted using "Method of least squares". This methodology is applicable only when response variable, viz. weight of a fish is assumed to be described by a single fixed number corresponding to each fixed value of explanatory variable, viz. length of a fish. But in reality, due to fuzziness or vagueness.of underIying phenomenon, weights of fish of exactly equal lengths need not be same. Thus, response vai'iable is fuzzy or vague as it can be described only by an interval. To deal with this situation, "Method of fuzzy least squares" is discussed. Finally, the methodology is illustrated on some length-weight data of fish culled from literature.