The existing condition-based maintenance methods based on inspection activities were only suitable for short-term schedules and involved only imperfect inspection or maintenance. A long-term optimization model was proposed based on on-site features of system maintenance scheduling by incorporating deterioration process with both imperfect maintenance and inspection activities. The delay time theory was introduced to describe deterioration process of system with double contingencies from imperfect maintenance and inspection. The proposed recursive reliability model can accurately evaluate the deterioration rate and failure rate of system. A maintenance scheduling optimization model was applied to minimize the average cost by searching appropriate inspection cycles and replacement strategy. The reliability and availability constraints were considered to meet the requirement of maintenance. Case studies show that the proposed model can attain optimal inspection strategy with minimum system cost. The accuracy of inspection was crucial for the scheduling optimization model compared with age-based and reliability-based maintenance models. The proposed model with a high accuracy of inspection reached the minimum cost compared to models without inspections.