Background and Objectives: Hypertensive disorders of pregnancy (HDP), including gestational hypertension (GHT) and preeclampsia (PE), share clinical features but differ in pathophysiology and risk profile. Pancreatic stone protein (PSP), a pancreas-derived acute-phase glycoprotein, has emerged as a biomarker of systemic stress. We investigated whether PSP could discriminate PE from both GHT and normotensive pregnancy, and whether this discrimination extends beyond the discriminatory information captured by conventional CBC-derived inflammatory indices. Materials and Methods: In this prospective case-control study, 84 pregnant women were enrolled across three groups: normotensive controls (n = 42), GHT (n = 20), and PE (n = 22). Serum PSP was measured by ELISA. Seven CBC-derived inflammatory indices (NLR, PLR, MLR, SII, SIRI, PIV, AISI) were calculated. Three-group comparisons used Kruskal-Wallis or ANOVA with appropriate post hoc testing. Receiver operating characteristic (ROC) analysis was performed across multiple clinical scenarios. Binary and multinomial logistic regression analyses were conducted with control as the reference category. Results: Median serum PSP levels showed a progressive elevation across the HDP spectrum: 6.91 (5.47-9.73), 9.08 (8.29-10.36), and 11.07 (10.03-11.83) ng/mL for control, GHT, and PE groups, respectively (Kruskal-Wallis p < 0.001). All three pairwise comparisons remained significant after Bonferroni correction (control vs. GHT, p = 0.035; control vs. PE, p < 0.001; GHT vs. PE, p = 0.002). PSP showed moderate-to-good exploratory discriminatory performance for PE versus control (AUC 0.83; sensitivity 95.5%, specificity 66.7% at 8.61 ng/mL), PE versus GHT (AUC 0.80; sensitivity 68.2%, specificity 85.0% at 10.70 ng/mL), and PE versus all non-PE participants (AUC 0.82). None of the inflammatory indices reached statistical significance (all p > 0.05), although several (MLR p = 0.072; SIRI p = 0.071) showed borderline upward trends consistent with the severity gradient. Multinomial logistic regression showed a consistent graded association: each 1 ng/mL increase in PSP was associated with a 42% increase in the odds of GHT (vs. control) (OR 1.42; 95% CI: 1.05-1.92; p = 0.022) and approximately threefold higher odds of PE (vs. control) (OR 2.99; 95% CI: 1.57-5.68; p = 0.001). Conclusions: Serum PSP demonstrated a graded increase across the HDP spectrum and showed moderate-to-good discriminatory performance for PE, including differentiation from GHT. These findings suggest that PSP may capture biological information not reflected by conventional CBC-derived inflammatory indices. However, the proposed cutoffs should be regarded as exploratory, and larger multicenter studies with gestational-age-matched controls and direct comparison with angiogenic biomarkers are required before clinical implementation.