Aim: Congestive heart failure (CHF) is a severe cardiovascular disorder seen in the Emergency Department (ED). B-type Natriuretic Peptide (BNP) is usually ordered to evaluate the CHF severity. However, it is difficult to interpret serum BNP level when different clinical entities existed. The aim of this study is to illustrate the correlation between serum BNP level and relevant clinical variables and further determine the role of serum BNP in different CHF patients. Methods: Univariate comparisons between 26 clinical variables and serum BNP level were analyzed. In order to avoid confounding factors, potential independent clinical variables were analyzed together using multivariate regression. Results: 529 CHF patients were reviewed and divided into different groups by 26 clinical variables. Serum BNP levels were found statistically significant different by univariate compareson between groups divided by 8 clinical variables that included obesity, diastolic/systolic heart failure (HF), serum blood urea nitrogen (BUN) level, serum creatinine (Cr) level, serum sodium (Na) level, patients taking loop diuretics, history of cerebrovascular accident (CVA), and history of dementia. Among all 8 clinical variables, obesity, serum BUN, Cr level, and diastolic/systolic HF had weak-to-moderate correlation effects with serum BNP level by correlation coefficient analysis. However, only obesity and dia- stolic/systolic HF were two moderately stronger clini- cal variables that can affect the serum BNP levels by multivariate regression. Analyzing CHF patients separately by obesity and diastolic/systolic HF subsets showed longer hospitalization in diastolic HF patients with relatively higher serum Cr level. In addition, poor correlation was found between serum BNP level and length of hospitalization (LOH) as well. Conclusion: High variability of serum BNP levels exists in CHF patients with weak-to-moderate correlation effects particularly on obesity and diastolic/systolic HF.It is recommended that physicians should be cautious on interpreting BNP in different CHF populations