Accurate　and　efficient　parameter　extraction　of　PV　models　from　I-V　characteristic　curves　is　significant　for　modeling,　evaluation　and　fault　diagnosis　of　PV　modules/array　s.　Recently,　a　large　number　of　algorithms　are　proposed　for　this　problem,　but　there　are　still　some　issues　like　premature　convergence,　low　accurate　and　instability.　In　this　paper,　a　new　improved　shuffled　complex　evolution　algorithm　enhanced　by　the　opposition-based　learning　strategy　(ESCE-OBL)　is　proposed.　The　proposed　algorithm　improves　the　quality　of　the　candidate　solution　by　the　opposition-based　learning　strategy.　Moreover,　the　basic　SCE　algorithm　evolves　with　the　traditional　competition　complex　evolution　(CCE)　strategy,　but　it　converges　slowly　and　is　prone　to　be　trapped　in　local　optima.　In　order　to　improve　the　exploration　capability,　the　complex　in　the　basic　SCE　is　evolved　by　a　new　enhanced　CCE.　The　ESCE-OBL　algorithm　is　compared　with　some　state-of-the-art　algorithms　on　the　single　diode　model　(SDM)　and　double　diode　model　(DDM)　using　benchmark　I-V　curves　data.　The　comparison　results　demonstrate　that　the　proposed　ESCE-OBL　algorithm　can　achieve　faster　convergence,　stronger　robustness　and　higher　efficiency.　(C)　2019　The　Authors.　Published　by　Elsevier　Ltd.