System engineering is the process of classifying the system requirements, organize them and establishing a common system architecture. E - service system architecture is the e - services system, structure or structures that include elements of the system components, characteristics and their interactions, which are required for e – service execution. For system design classically is used two design methods - a functionally-based and object-oriented. In development of e - services it is important to select e - service system architecture to meet all the set of quality criteria to promote the usefulness of e - service. This means that it is necessary to use effective methods of designing e - services systems that provide an acceptable e - services system architecture creation. In such cases, a simple functional-based or object-oriented design approach is inadequate. Methods that address quality system issues in the system design phase for e – service system architecture design can be used. Such as Attribute- Driven Design method ADD, the Software Architecture Analysis method SAAM, Jan Bosch software architecture design method, Architecture Tradeoff Analysis Method ATAM, Hazard Analysis of Software Architectural Designs HASARD, etc. These methods address the question of the quality, but they have one major drawback. Systems architect or a person carrying out the system architecture design, subjective opinion, as well as the possibility that he will be missed by far the best solution, or it will not be discussed. The solution can be selected, that is not the best in case and can be only local minimum because not all possible cases are discussed. In such cases, the e - services system design method can be used, which is based that web services are designed on the multicriteria optimization. The proposed method consists of seven major design steps. In the beginning it is necessary to create the e-service algorithm graph that describes the e-service to be designed. In the e-service algorithm graph, it is necessary to determine the possible restrictions for inter-segmentation of vertexes. By recursively segmenting the e-service algorithm graph, the web service graphs are obtained that can be used as the basis for the development of the e-service system architecture. The e-service algorithm graph is assumed as the initial web service graph. The numerical vales of quality metrics of the obtained web service graphs are calculated using the calculation formulae. Then from all the web service graphs, the Pareto optimality set is obtained and the web service graphs of the obtained set can serve as the basis for selecting an acceptable e-service system architecture. The web service graphs that are contained in the obtained set can be designed in detail and implemented in the e-service system architecture In most cases, the Pareto optimality set contains more than one solution. If the obtained set is sufficiently small, the selection of web service graphs from the set can be left to the system designer. However, if there are many possible solutions, then in order to find the particular web service graph, the criteria can be decreased or combined with another optimisation or design method, e.g. weighted method or Jan Bosch method. The described design methods are mutually compared that the system architect can choose applicable design method option for e – service system architecture design.