We present a quantitative method based on stochastic Petri net (SPN) to evaluate the survivability of distributed service. Three types of combination service models: one-to-one, one-to-N, and M-to-N are studied, which can reflect survivability of distributed service from a different perspective. Absorbing-state reliability, maintainability and steady-state availability are considered as measures of survivability, and calculated by reduced SPN models and underlying Markov chains (MC). In terms of this method, we make a survivability evaluation systematically by taking Finance Service Network (FSN) as an example. Numerical results indicate that the approach of this paper is effective and efficient for evaluation the survivability of distributed service.