Background: Acute pancreatitis (AP) can be life-threatening with unpredictable severity. Despite advances in management, its pathogenesis remains unclear. This study investigated metabolites and lipoprotein profiles in AP patients of African descent to understand the underlying pathophysiological conditions so as to inform prognosis and management. Methods: Serum samples were collected from 9 healthy controls (HCs) and 30 AP patients (8 with mild AP, 14 with moderately severe AP, and 8 with severe AP) on days 1, 3, 5, and 7 post epigastric pain and subjected to nuclear magnetic resonance (NMR) spectroscopy. Wilcoxon and Kruskal–Wallis rank-sum tests compared numerical covariates. Lipoprotein characterization was performed using the Liposcale test, and Spearman’s rank test assessed data correlations. The p-values < 0.05 indicated significance. Results: Thirty-eight metabolic signals and information on lipoprotein subclasses were identified from the NMR spectra. The severity of AP correlated with increased levels of 3-hydroxybutyrate and acetoacetate and decreased levels of ascorbate. Distinct metabolic phenotypes were identified and characterized by unique inflammatory and lipoprotein profiles. High-density lipoprotein cholesterol (HDL-C) decreased across all the metabolic phenotypes of AP when compared with the HC, while elevated immediate density lipoprotein cholesterol (IDL-C) and very low-density lipoprotein cholesterol (VLDL-C) levels were observed. Time-dependent changes in metabolites were indicative of responsiveness to therapy. Conclusions: Our findings indicate that dysregulated metabolites and lipoproteins can be used to differentiate AP disease state and severity. Furthermore, integrating clinical parameters with data on metabolic and lipoprotein perturbations can contribute to a better understanding of the complex pathophysiology of AP.