In order to better simulate the condition states of the system and get a more accurate reliability of the system, this paper investigates three failure modes for a single component system: Multi-state degradation and two failures due to extreme shocks and cumulative shocks. A comprehensive consideration of several internal and external factors for a system with multiple condition states would make the model more practical, applicable and realistic compared with the previous researches which only consider one side. Different degradation rates and shock rates are used simultaneously in one model, considering the different states of the system. Deterioration level and effects of shocks, corresponding system reliability is established according to the probabilistic and statistics theory. The sensitivity analysis of variables in system reliability model is also conducted to identify the most significant factors. Our results show that: (1) The system reliability decreases with accelerating rate and then the decelerating rate after reaching the peak. (2) Reliability is significantly sensitive the state transition parameters, internal degradation means and intensity of external shocks, where the company would benefit a lot from finding the optimal value of these variables, and improve the reliability of the whole system.