This paper describes a reliability based study of OCTG connections and casing for thermal wells. The fundamental principle of structural reliability analysis is being used to model the uncertainties associated with the load capacity of the casing as well as different installation and the operational loading conditions. These uncertainties influence the strength and performance of the thermal well casing structure. Traditional deterministic design factor cannot quantify the uncertainties associated with these variables. Statistical distributions of these variables obtained from different sources such as analysis, production and laboratory test data. Finding the proper statistical distribution of installation and operational load of the represented thermal well data is challenging since each well is unique. However, with the reliable dataset of statistical significance used to build a structural reliability model with reasonable accuracy. This model used for risk analysis as well as could predict the failure mode during the early design phase. The probability theory used in the assessment of structural reliability model. The reliability model concerns the aspects of probabilistic modeling of load (S) and strength/resistance (R) variables. The combined loading conditions and elastic-plastic materials models are outlined to represent the thermal well. Structural reliability analysis methods such as FORM/SORM and Monte Carlo simulation will be discussed to address the randomness of casing loading and OCTG property uncertainty originating from the parameters mentioned above.