Reliability allocation is a critical step during the early design and development stage of a complex system. The majority of existing methods allocate reliability to individual subsystems considering diverse factors, and fail to recognize the non-linear phenomenon for failure rate to obtain the potential for improvement. The paper proposes a minimum improvement effort based reliability allocation method for allocating system level reliability improvement target to environmental stresses that subsystems will suffer. By fuzzy allocation and AHP, weightage factors of subsystems and stresses can be respectively acquired. As improvement effort is exponential with failure rate, the reliability growth effort function is modeled. With weightage factors and effort function, we put forward the objective function with minimizing the system loss. The method takes system's reliability as constraints, and models the minimum improvement effort based reliability growth allocation method. The result can be obtained effectively with lingo. A case example is presented to demonstrate the effectiveness of the proposed approach.