This study utilises Accelerated Failure Time (AFT) models to assess survival data from HIV/AIDS patients in Sikkim provided by the Sikkim AIDS Control Society (SACS). The analysis investigates four parametric distributions—Weibull, Log-Normal, Exponential and Log-Logistic—to find the best fit for the data. The main objective is to discover significant characteristics that affect patient survival, such as age, gender, co-morbidities, education and socioeconomic position. The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were employed in selecting models, with the Log-Normal distribution proving to be the best fit. The results show that co-morbidities, age and educational attainment have a major influence on survival outcomes. The robustness of the chosen model was validated by model diagnostics such as residual analysis and validation against Kaplan–Meier survival estimates. These findings create a strong methodological foundation for survival analysis in epidemiological research and offer important new information about the factors influencing HIV progression in Sikkim.