Seismic risk calculation is based on stochastic models describing at best the seismic activity. The focus is on calculating earthquakes occurrences probability and analyzing spatiotemporal evolution of these probabilities. The point process theory is the main tool for modeling seismic data and taking into account past history prior to a given date. Thus, probability and statistical tools were developed by scientists from research teams located in regions with very high seismicity regions: California (Turcotte), Japan (Ogata) and New Zealand (Vere-Jones). Statistical inference is based on likelihood techniques in order to optimize information processing and projection strategy. We consider the ETAS model and analyze the epicenter and magnitude distribution of earthquakes occurred in the Lesser Antilles arc from 1999 to 2004. A comparison of different probability density functions for the triggered event times is performed.