This paper presents an evaluation of Bayesian Net- work (BN) performance using different elimination order heuris- tics in modeling the relationship between occupational strain (OS) and coronary heart disease (CHD). The study compares three elimination strategies—random order, min-order, and min- fill heuristics—across MAP (Maximum A Posteriori) and MPE (Most Probable Explanation) algorithms. An experiment was conducted with varying Bayesian network sizes to assess run time, number of nodes, and tracked degrees. Results indicate that min- order and min-fill heuristics significantly outperform the random order in terms of efficiency, especially with larger networks. The Bayesian Network model incorporates variables such as gender, job type, mobbing, job demands, income, social opportunities, and various health factors to explore their interplay in OS and CHD. Analysis of prior and posterior marginals, as well as MPE and MAP queries, provides insights into the likelihood of CHD given different conditions. Key findings include the higher likelihood of females reporting CHD, the counterintuitive result that heavy smokers with high job demands are less likely to develop CHD, and the identification of men with active job types as most likely to experience OS. Additionally, d- separation tests show that OS and mobbing are independent of gender and job type in the given context. This study highlights the effectiveness of elimination heuristics in BN reasoning and underscores the complex interrelationships between job strain and health outcomes.