The invention discloses a multi-objective optimization solving method for cloud instance selection. The multi-objective optimization solving method comprises the following steps: A, background problemdefinition: defining a problem background of a solver; b, generation of an initial complete Pareto set: obtaining the initial Pareto set with the demand number of 1; c, generalized Cartesian productgeneration: carrying out generalized Cartesian product on the complete Pareto set and the initial Pareto set; d, generation of an intermediate complete Pareto set: performing traversal operation on the generalized Cartesian product to obtain a new complete Pareto set; and E, final complete Pareto set generation: obtaining a final complete Pareto set through recursive C and D operations. Accordingto the method, the multi-objective optimization problem similar to a cloud instance type selection problem can be solved, a complete Pareto set can be obtained, and a user is helped to make a decisionmore accurately. In addition, the similar multi-objective optimization problem can be solved, and a complete Pareto set is obtained.