In the construction industry, choosing a different architecture will lead to very different techniques, costs, and project implementation time. The same goes for software projects. Choosing an appropriate software architecture requires consideration of many factors. This choice significantly affects the success rate of the project. Many software architectures have been proposed in the past and have been successfully applied in the software industry. This study proposes a model to evaluate the influence of software architecture on the accuracy of software effort estimation. This work is done by clustering data according to software architecture and then applying the FPA method and the proposed model to estimate software effort. Experimental results show that clustering data according to software architecture provides higher estimation accuracy than non-clustering data on both the FPA method and the proposed model.