Falls are one of the very serious problems in the healthcare system for the elderly, often resulting in a rapid decline in functionality and death. Serious consequences of sustaining a fall include broken or fractured bones, superficial cuts and abrasions as well as soft tissue damage. Several solutions are proposed to resolve such problems, however, major difficulties they encounter are cost, comfort and performance. In this paper, we propose a new Android application for fall detection based on theory of machine learning and data mining. Concretely, we used a classification method to detect if it is a fall down or another common events (e.g. sitting, jumping, etc.). The obtained experimental results show that our proposal yielded impressive results compared to related works.