Probabilistic　flexural　capacity　models　and　probabilistic　calibration　method　for　corroded　reinforced　concrete　(RC)　columns　under　both　large　and　small　eccentric　compression　were　established　by　combining　with　theoretical　analysis　and　209　sets　of　test　data.Expressions　of　probabilistic　flexural　capacity　for　corroded　RC　columns　under　both　large　and　small　eccentricity　were　derived　by　taking　into　account　the　equilibrium　condition　of　internal　force,　the　influences　of　corrosion　on　material　and　bond　properties　of　steel　bar　and　concrete,　and　the　uncertainties　of　probabilistic　model　parameters　and　model　error.　The　posterior　information　of　probabilistic　model　parameters　was　obtained　by　the　Bayesian　theory　and　Markov　Chain　Monte　Carlo　(MCMC)　method.Prediction　accuracy　of　traditional　deterministic　models　were　calibrated　based　on　the　probability　density　function　(PDF),　confidence　interval　and　confidence　level.Analysis　results　show　that　the　probabilistic　flexural　capacity　models　have　high　accuracy　and　it　describes　well　the　probabilistic　characteristics　of　flexural　capacity.　The　accuracy　of　traditional　deterministic　models　for　large　eccentric　compression　was　relatively　high,　but　the　accuracy　of　traditional　deterministic　models　for　small　eccentric　compression　was　relatively　low　and　with　large　discreteness.　©　2022,　The　Editorial　Board　of　Journal　of　Basic　Science　and　Engineering.　All　right　reserved.