Breast cancerBreast cancer causes a significant number of annual deaths and is the principal type of cancer that results in the deaths of women worldwide. Developing cancer predictionPrediction and diagnosis is crucial for maintaining a healthy life. It is imperative to achieve high accuracy in cancerCancer prediction to improve treatment options for patients and increase their chances of survival. HealthcareHealthcare datasets are readily accessible, encouraging researchers to use them for study. So, many researchers work on it, and limitations still exist. In this work, a performance comparison between an EnsembleEnsemble and a proposed Convolutional Neural Networks (CNNs)Convolutional Neural Network (CNN) Models is conducted to overcome the shortcomings of single-model performance on breast cancerBreast cancer predictionPrediction. Also, we have applied data balancing techniques to get more accurate results. The results found that the proposed CNN Model had a higher level of accuracy, reaching 97.65%. In comparison, the EnsembleEnsemble ModelEnsemble models achieved an accuracy of 96.21%. Additionally, the proposed CNNConvolutional Neural Network (CNN) Model demonstrated a precision of 95.21%, a recall of 96.43%, and an F-score of 96.35%, while the EnsembleEnsemble ModelEnsemble models had a precision of 94.13%, a recall of 95.76%, and an F-score of 95.50%. The proposed CNN Model presented superior accuracy and a more even balance between precision and recall, proving its effectiveness in detecting positive breast cancerBreast cancer cases, and the goal is to reduce differences in diagnoses and improve accuracy by merging these models.