After elaborating the theory of the genetic algorithm, including the coding,initial population, fitness function, genetic operation and controls parameter of arithmetic, this paper presents its application to feed diet of fishing in detail. It adopts real coding, linear ranking selection, arithmetical crossover and nonconforming mutation. The results show that this algorithm is better than the optimization of common application mathematics in the feed diet of fishing. It mainly embodies that the unitage cost price of the former is lower than that of latter. Due to resolving the long degree non-linear problem with more factors, the diet by the genetic algortithm more well and truly accords with the nutrition tontent criterion of corresponding fish.