Landslides represent a serious hazard in many areas around the world, potentially leading to human losses and significant damages to structures and buildings. For this reason, over the years a consistent number of studies and researches have been carried out to analyse these natural phenomena and their evolution. This study presents the application of an automatic procedure specifically developed to identify the onset of landslide acceleration by analysing monitoring displacement data with a multi-criteria approach. The proposed procedure aims to identify this point by applying a four-level validation process on a pre-determined dataset. Once the analysis returns a positive result for a certain number of monitoring data, it is possible to state that the landslide reached the accelerating phase of its evolution, thus allowing to define a specific point representing the onset of acceleration. The method was applied to several historical case studies taken from scientific literature, in order to test its practicability and effectiveness. This procedure could be especially useful in Early Warning Systems where time of failure forecasting models are implemented, allowing to improve their performances by providing an automated and reliable procedure to define the beginning of potentially critical landslide events.