Background-Gestational Diabetes is a type of diabetes that generally happens in females during the time pregnancy. This diabetes causes significant health risks to mothers and fetuses. This diabetes can affect all the trimesters of pregnant women. It can also affect the Postpordinal time of pregnancy as well as the prepordinal time. Early prediction and management of this diabetes are crucial. Artificial Intelligence model appeared as a important tool for Gestational Diabetes Prediction. Methods-A systematic research is conducted by implementing the AI model for gestational diabetes predictions. The AI models were implemented by using the Gestational Diabetes Dataset from Iraq Kurdistan region laboratories, which collected datasets from pregnant women with and without diabetes. The three AI models used here for the prediction of gestational diabetes are Generative Adversarial Networks (GANs), Variational Autoencoder (VAE) and a hybrid model Vanilla GAN with Min Max scalar. Results-In the Gestational diabetes prediction process, the result is calculated on the metrics of the Confusion Matrix. Hybrid model Vanilla GAN calculates the highest accuracy with a Min-Max scalar of 85%. Whereas the Generative Adversarial Networks calculated, accuracy is 82% on the 125 epoch size. The Variational Autoencoder (VAE) calculated accuracy is 80% on the 125 epoch size. Conclusion-AI models for Generative Learning for Gestational diabetes prediction show accurate results. It provides the evolution of a diabetes prediction model in less time.