Introduction: The severity of COVID-19 disease can be viewed from the massive death rate recorded worldwide so far. The majority of increase in death rate is due to late identification of disease. Aim: To detect COVID-19 from Chest X-ray images using concatenated Deep Learning Neural Networks Xception with ResNet152V2 and Xception with EfficientNet-B7. Materials and Methods: This work on Deep Learning (DL) system proposes the concatenation of two DL networks to identify COVID-19 using X-ray images. They are Xception with ResNet152V2 and Xception with EfficientNet-B7. Initially, the input X-ray images are performed with pre-processing. The pre-processed images are given to Xception with ResNet152V2 or Xception with EfficientNet-B7. Various features are extracted from these two networks. The output features from Xception and ResNet152V2 or EfficientNet-B7 are concatenated. The concatenated features are then given to the classifier for the classification of COVID-19. Results: The implementation has been performed on Google Colab using the neural networks with Keras library with a usage of upto 12.69 GB RAM. The average accuracy for COVID-19 is 62% and 60% using concatenated Xception with EfficientNet-B7 and concatenated Xception with ResNet152V2 respectively. Conclusion: The proposed concatenated nets provide better results for 15-epoch with a batch size of 5. With an increase in epoch and batch size the accuracy of the proposed method will be increased upto 99.7%.