To address the challenge of extracting fault features from rolling bearings, especially when early signals are weak, this study introduces a fault diagnosis framework for rolling bearings utilizing an enhanced Black Wing Kite algorithm (IBKA) to optimize the Variational Mode Decomposition (VMD) for feature extraction, coupled with a Support Vector Machine (SVM) for classification. Initially, the IBKA is employed to fine-tune VMD parameters, utilizing minimum envelope entropy as the objective function to determine the optimal parameter set (K, α). Subsequently, the fault signal is decomposed by VMD based on the identified parameters, and the intrinsic mode function (IMF) component with a higher kurtosis value and a lower envelope entropy is selected as the optimal signal component. Conclusively, a feature vector is assembled from the nine time-domain indicators of the chosen IMF component and fed into the SVM for training and fault diagnosis. The findings indicate that the IBKA-VMD model achieves a fault diagnosis accuracy of 99.33%, outperforming the BKA-VMD model by 2.33%. This model is adept at feature extraction and classification, offering technical assistance for the early detection of rolling bearing faults.