The relevance of reliability evaluation strongly depends on the quality of input data as failure rates. Reliability data handbooks give generic values which do not often fit system specificities. This paper deals with influencing factors in order to take into account some aspects as design, environment and use in reliability evaluations. Once a definition and a classification are proposed, a brief review of existing models is presented. This paper also introduces a new failure rate evaluation with influencing factors especially developed for safety instrumented systems. The seven-step methodology combines both qualitative and quantitative analyses to compensate for a potential lack of feedback knowledge. Some criteria are used to set a failure rate within a prior interval, according to system conditions. An application regarding safety pressure relief valves is included. The expected better argued and accurate results aim at leading to a more efficient risk management.