Objective To explore the value of constructing a nomogram based on the combined X-ray and MRI imaging characteristics for predicting the risk of early non-traumatic osteonecrosis of the femoral head （NONFH） collapse. Methods Clinical， X-ray， and MRI data from 149 early-stage NONFH patients （237 hips） at two medical centers were collected and then divided into training and validation cohorts. Univariate and multivariate Cox regression analyses were carried out to pinpoint independent predictors of femoral head collapse. Single-factor and multi-factor predictive models for 3-year and 5-year outcomes were developed. The performance of the models was evaluated using the area under the curve （AUC）， and the consistency of the models was assessed using the concordance index （C-index）. The optimal model was chosen for nomogram construction. Internal validation was carried out with the training cohort， and external validation was performed using the validation cohort. Results A total of 159 hips from center 1 and 78 hips from center 2 were incorporated into the study. Multivariate Cox regression analysis pinpointed the JIC classification （HR = 37.78， 95%CI： 4.631 - 308.15， P < 0.001）， joint effusion （HR = 1.613， 95%CI： 1.037 - 2.509， P = 0.034）， and necrosis index （HR = 1.016， 95%CI： 1.003 - 1.028， P = 0.012） as independent predictors of femoral head collapse. Both internal and external validations verified that the combined model for 3-year and 5-year prediction displayed superior performance， enhanced accuracy， and increased net benefit. Calibration curves revealed a good concordance between the risk probabilities predicted by the nomogram and the actual observed outcomes. Conclusion The nomogram constructed on the basis of X-ray and MRI imaging features can effectively predict the risk of femoral head collapse in patients with NONFH， offering crucial guidance for the formulation of individualized clinical intervention strategies.