The health-science driven datasets present several common challenging characteristics for their analysis, notably the presence of censored observations and highly correlated covariates. When data variables exhibit strong correlations, bootstrap samples from the prior distribution can obscure individual effects, posing a challenge to accurately discern the true relationship between covariates and the response variable. This phenomenon can result in unrealistic estimates, underscoring the importance of incorporating the correlation structure into the modeling process. Therefore, analyzing survival data requires specialized techniques tailored to handle these distinctive characteristics. The impact of censored data observations and the possible solutions to this problem have been well studied within survival models, but there remains a significant gap in exploring methodologies that effectively account for the multidimensional distribution structure of highly correlated variables. This article aims to apply the Proper Bayesian bootstrap, proposed by Muliere and Secchi, used in sampling the posterior distribution over ensemble trees incorporating the prior distribution of highly correlated variables for analyzing survival data. The model’s performance is assessed through a simulated study, and the results are compared to traditional survival models, such as the Cox model and Survival Random Forest, in terms of the average Brier score, particularly with varying sample sizes.