This　paper　is　devoted　to　developing　an　efficient　and　robust　stress　updating　algorithm　in　a　relatively　simple　computational　framework　to　address　the　two　difficulties　of　implementing　elastoplastic　soil　models,　namely　the　nonsmoothness　and　the　nonlinearity.　In　the　proposed　algorithm,　the　nonsmoothness　caused　by　the　loading/unloading　inequality　constraints　is　eliminated　by　replacing　the　Karush-Kuhn-Tucker　conditions　with　the　smoothing　function.　The　stress　updating　can　be　achieved　by　solving　a　set　of　smooth　nonlinear　algebraic　equations　in　this　algorithm.　The　nonlinear　equations　are　solved　using　the　line　search　method,　which　allows　a　larger　convergence　radius　of　the　solution　in　contrast　to　the　standard　Newton　method.　Meanwhile,　the　smoothing　consistent　tangent　operator　corresponding　to　the　unconstrained　stress　updating　strategy　ensures　the　quadratic　convergence　speed　of　the　global　solution.　The　modified　Cam-clay　model　is　used　as　an　example　to　demonstrate　the　implementation　of　this　algorithm.　The　correctness,　computational　efficiency,　and　robustness　of　the　algorithm　are　validated　and　assessed　by　comparing　it　with　the　analytical　solutions　in　case　of　cylindrical　cavity　expansion　and　the　ABAQUS/Standard　default　integration　method.　In　simulations　with　large　load　increment　sizes,　the　CPU　time　consumed　by　the　new　algorithm　can　be　less　than　half　of　the　ABAQUS　default　algorithm.