Although the mobile AR application is getting popular, only a limited number of researches are available. As mobile AR application become more popular, it is obvious that the efficient in interaction for mobile environments are necessary. However, precise interaction with touch screen interfaces is difficult in mobile AR environment due to the shaky viewpoints even though gyroscope and accelerometer is already built-in. Currently, interaction using touchscreen have become the de-facto standard which limit the interaction between user and devices.This paper in turn presents a multimodal interaction model for mobile AR application. The design guideline was developed based on literature review from past research papers. The findings indicate that there are 3 levels of requirement in multimodal interaction and several important key elements of selecting the modalities as the first step in creating a mobile AR application. © 2014 WIT Press.