<p>To investigate the correlation between serum endocrine hormone levels and the malignancy degree of prolactinomas, and analyze their predictive value for patient prognosis. A total of 100 prolactinoma patients admitted to the Affiliated Hospital of Xuzhou Medical University from January 2019 to December 2024 were enrolled. Based on tumor invasiveness, patients were divided into benign (<i>n</i> = 74) and malignant (<i>n</i> = 26) groups. Serum endocrine hormone levels were compared between groups. Pearson’s test analyzed correlations between hormone levels and tumor malignancy. According to new metastases, recurrence, or death during follow-up, patients were classified into good prognosis (<i>n</i> = 69) and poor prognosis (<i>n</i> = 31) groups. Multivariate logistic regression identified factors influencing poor prognosis. Restricted cubic spline analysis evaluated dose-response relationships between hormone levels and poor prognosis risk. A nomogram model was constructed and its predictive performance evaluated. The malignant group showed significantly higher serum prolactin (PRL) levels but lower free thyroxine (fT4) levels than the benign group (<i>P</i> &lt; 0.001). Serum PRL positively correlated with tumor malignancy (<i>r</i> = 0.460, <i>P</i> &lt; 0.001), while fT4 showed negative correlation (<i>r</i>= − 0.453, <i>P</i> &lt; 0.001). Multivariate analysis revealed giant tumor type and elevated PRL as risk factors for poor prognosis, whereas pseudocapsule presence and increased fT4 were protective factors (<i>P</i> &lt; 0.05). Nonlinear relationships existed between poor prognosis risk and PRL/fT4 levels (Pnonlinear &lt; 0.05). In the nomogram model, all four factors had variance inflation factors (VIF) &lt; 5 (1.043–1.091). The model’s ROC curve area was 0.888, and Hosmer-Lemeshow test confirmed good accuracy (χ² = 12.673, <i>P</i> = 0.124). Serum PRL and fT4 levels significantly correlate with prolactinoma malignancy degree and influence patient prognosis.</p>