In this paper, main characteristics and peculiarities of service tasks are investigated in order to propose design criteria in the form of computationally efficient objective functions. Then, these design criteria can be implemented in a multi-objective optimization algorithm by using ever commercial packages in order to obtain optimal design solutions for service robots. The proposed procedure has been applied to a robotic arm for service tasks that is composed of two modules having altogether four degrees of freedom (dofs). The main module with two dofs is called as SIDEMAR (Spanish acronym of Mechatronic Design of an Integrated System for Service Robots). It has been designed and built at Carlos III University of Madrid as a module for service robot applications. Results of the above-mentioned case of study are reported to show both the soundness and engineering feasibility of the proposed multi-objective optimal design procedure.