Accurate,　efficient　and　reliable　parameter　extraction　of　solar　photovoltaic　(PV)　models　from　the　measured　current-voltage　(I-V)　characteristic　curves　is　important　for　evaluation,　modelling,　and　diagnosis　of　the　actual　operating　state　of　in-situ　PV　arrays.　In　recent　years,　numerical　heuristic　optimization　algorithms　based　parameter　extraction　methods　have　been　proposed.　However,　the　efficiency　and　reliability　of　these　methods　are　limited　due　to　heuristic　or　stochastic　searching　strategies.　In　this　paper,　by　combining　the　trust-region　reflective　(TRR)　deterministic　algorithm　with　the　artificial　bee　colony　(ABC)　metaheuristic　algorithm,　a　new　hybrid　algorithm　ABC-TRR　is　proposed　to　improve　the　parameter　extraction　of　PV　models.　The　ABC-TRR　algorithm　combines　the　global　exploration　capability　of　the　ABC　and　the　local　exploitation　of　the　TRR,　which　achieves　a　good　tradeoff　among　accuracy,　convergence　and　reliability.　The　proposed　ABC-TRR　hybrid　algorithm　is　evaluated　and　compared　with　other　state-of-the-art　algorithms　using　the　standard　I-V　curves　of　the　benchmark　Photowatt-PWP201　PV　module　and　RTC　France　solar　cell　as　well　as　the　measured　I-V　curves　of　a　laboratory　PV　module/string/array.　Comprehensive　experimental　analysis　and　comparison　results　demonstrate　that　the　proposed　ABC-TRR　algorithm　achieves　the　same　level　of　accuracy　as　the　best　reported　algorithms　with　the　highest　overall　reliability.　More　importantly,　the　ABC-TRR　algorithm　converges　4.69　times　faster　than　the　best-reported　algorithms　on　average.　In　view　of　these　advantages,　the　proposed　ABC-TRR　algorithm　is　a　promising　altemative　for　accurately,　efficiently　and　reliably　extracting　the　parameters　of　PV　models　from　measured　I-V　curves.　In　addition,　it　was　experimentally　demonstrated　that　the　parameter　extraction　result　can　be　used　to　indicate　the　partial　shading　and　abnormal　degradation　conditions.