Nowadays there is growing awareness of the importance of Systems of Systems (SoS) which are large-scale systems composed of complex systems. SoS possess specific properties when compared with monolithic complex systems, in particular: operational independence, managerial independence, evolutionary development, emergent behavior and geographic distribution. One of the current main challenges is the impact of these properties on SoS security modeling and analysis. In this research proposal, we introduce a new method incorporating a process, a language and a software architectural tool to model, analyze and predict security architectural alternatives of SoS. Thus security will be taken into account as soon as possible in the life cycle of the SoS, making it less expensive.