To　further　enhance　energy　production　and　reduce　the　Levelized　cost　of　energy　(LCOE)　it　is　essential　to　plan　spare　parts　to　reduce　the　downtime　of　the　photovoltaic　(PV)　systems　with　micro-inverter　(MI)　caused　by　the　failure.　The　present　investigations　only　consider　the　deterministic　effects　of　spare　parts　of　PV　products　on　energy　yield.　The　spares　optimization,　especially　considering　the　stochastic　inventory　policy,　for　the　PV　system　has　not　been　reported　yet.　In　this　paper,　an　optimization　model　of　generating　capacity　incorporating　inventory　policies　of　spare　parts　is　constructed　for　the　MI　PV　system　combining　the　Continuous　time　Markov　chain　(CTMC)　with　stochastic　inventory　theory.　The　range　of　the　installation　scale　from　100　to　2000　is　investigated　to　analyze　the　effect　of　spare　parts　on　the　energy　yield.　We　find　that　the　installation　scale　being　more　than　200　for　a　500　W　PV　system　should　reserve　spare　parts.　At　the　installation　scale　approaching　2000　PV　systems,　the　LCOE　decreases　by　0.0063　$/kWh,　about　15　%.　Compared　to　not　stocking　the　spare　parts　of　policy　(-1,1),　the　cost　saves　8.3　x　10　boolean　AND　5　$,　more　than　14　%.　The　optimal　profits　deduced　by　the　proposed　methodology　can　enhance　the　competitiveness　of　the　utilities　and　the　PV　manufacturer.