Objectives <p>Pertussis is an important contributor to respiratory morbidity and mortality and remains underdiagnosed since pertussis re-emergence becomes a global public health concern in recent years. This multi-center retrospective observational study aimed to develop a diagnostic prediction model based on laboratory blood parameters to identify children with pertussis.</p> Methods <p>A total of 545 children with suspected pertussis from the hospital between January 2024 and June 2024 were identified and randomly split into training (<i>n</i> = 381), and validation group (<i>n</i> = 164). A model was generated based on Least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis identifying potential predictors of pertussis. Model performance was assessed using receiver-operating characteristic (ROC), calibration plots and decision curve analysis (DCA). External validation was conducted on an independent cohort from another hospital (<i>n</i> = 594).</p> Results <p>Based on LASSO and multivariate analysis, a model was formulated incorporating four predictors, including age, CRP level, lymphocyte count, and hematocrit (ROC: 0.741 for the training cohort, 0.694 for the internal validation cohort and 0.675 for the external validation cohort). Calibration curve and DCA validated the fitness and clinical application value of this nomogram.</p> Conclusions <p>The model developed in this multicenter cohort study exhibits high accuracy in predicting the probability of pertussis in children. This model can be very helpful for clinicians when making decisions.</p>