Early warning system aims to avoid and reduce the risk of damage and casualties caused by natural disasters. Landslides and earthquakes are natural disasters that often occur in Indonesia. The objective of the research is to develop landslide and earthquake early warning system applications for smartphone devices with Android operating system. Current smartphone contains low cost sensors able to detect movement. Despite of less accuracy than scientific instruments, but the ubiquity of smartphone can cover larger area. Crowdsourcing the data collection also reduces the false positive rate caused by the sensor. We embed our prediction model in Android application to distinguish between earthquake and common human activity. Further, the data is sent to cloud server if predicted as earthquake. When the number of user reported the same event at the same time in certain area met the threshold, push notification then delivered to smartphone user which may have impact by the event. We use data from Sistem Informasi Kebencanaan - Longsor by Research Center for Geotechnology to develop landslide early warning system. When smartphone user entering area with highly landslide potential, the would receive notification about the area potential. This research uses Agile method to develop early warning system on Android device. Research activities are conducted in 6 (six) phases: Requirement, plan, design, develop, release, track & monitor. In this paper, we described the process of design phase system that will be a guide in the next process in making the next program.