A particle swarm optimization with aging and challenging mechanism (ACM-PSO) is proposed for evaluating the capacity credit of photovoltaic (PV) generation from load side and generation side respectively. Firstly, on the basis of system state variables constructed by the number of types of conventional unit, ACM-PSO is adopted as a scan and classification tool to screen out the system failure state set which contributes to generation reliability indices. Secondly, proportional load increase mode is defined, and with the loss of load expectation remaining unchanged as the principle before and after the PV access, objection functions are constructed both from load side and generation side, and application of ACM-PSO for evaluating the effective load carrying capacity and equivalent conventional unit capacity of PV is executed. Finally, the proposed method is applied to an improved IEEE-RTS79 system to verify its effectiveness.