In public opinion, protection against asteroids impact has always been on the agenda of space engineering. Actually it started from 1994 when Shoemaker Levy stroke Jupiter. This protection works in two steps: detection of threat and deflection. Some space agencies and foundations monitor the sky and set up scenario. Although the sky is nowadays well monitored and mapped, there is no global plan nowadays against this threat. This paper focuses on the deflection step, and aims at forecasting which variables are involved and their consequences on the deflection mission. In fact the result depends on several factors, like the time before hazardous moment, the accuracy of detection tools, the choice of deflection method, but the most unpredictable are human factors. This study shows a strategy and so tries to give some new response parts to the global deflection problem.