The invention provides a pet feeding method and system based on pet daily data analysis. The method comprises the following steps: collecting pet species, gender, age, heartbeat frequency, blood pressure, body temperature, activity amount, feeding type, feeding amount, current image and current weight to form an influence factor matrix X, and uploading the data to a server, wherein the feeding type and the feeding amount form a decision variable; establishing complex nonlinear relation between the influence factor matrix X and pet health indexes in the server by utilizing an Elman neural network to obtain a pet feeding model; carrying out optimization on the pet feeding model by utilizing an NSGA-II algorithm to obtain a group of optimum solutions of the decision variable; serving the group of optimum solutions of the decision variable as recommended decisions X* of the pet and sending the recommended decisions through the server to a terminal device of a user for display; and feeding the pet by the user according to the recommended decisions X* displayed by the terminal device. The method and system can determine an optimal pet feeding scheme, and builds a better living environment for the pet.