Deep learning models’ exploration of synthetic data has been comparatively understudied compared to other areas of research. This paper proposes an assessment of discriminative deep learning (DL) models for classification task based on synthetic pandemic dataset to identify a robust and efficient model for a relevant application. In a specific term, it provided an empirical assessment performance of three different discriminative DL models: CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), and MLP (Multilayer Perceptron) for investigation. The study architecture depicts a complete classification and detection procedure of the presence COVID-19 pandemic synthetic dataset. It consists of four phases: (i) dataset collection and description phase, (ii) preprocessing phase; (iii) discriminative DL modelling phase; (iv) classification and evaluation phase. The research outcome shows that RNN achieved the highest classification report, it is observed that RNN achieved the highest classification report values of 0.9881 for recall, precision, F1 Score and accuracy metrics and a little higher ROC-AUC where it was measured 0.9995. The RNN model achieves perfect scores on all metrics, indicating an outstanding performance. This could be due to a variety of factors, such as the nature of the dataset.