The detrimental effects on human health caused by air pollution have made air quality a step of the utmost significance. The application of artificial intelligence is seen as promising. However, the challenges of applying machine learning and deep learning algorithms to regression problems make testing these models necessary. Moreover, the prediction accuracy of state-of-the-art models varies with different pollutants and is acceptable for certain pollutants only. This work is focused on multidimensional time series prediction using the Deep Convolution Generative Adversarial Network (DCGAN). The proposed methodology illustrates the reliability of the architecture with a “rolling window” approach for long-term prediction. Statistical metrics, namely R2 score, MSE, RMSE, and MAE, are used to evaluate the proposed work. It is then compared with traditional Machine Learning (ML) and state-of-the-art Deep Learning (DL) models. The R2 score of the model is found to be 0.99, with a sound train-to-test ratio of 4:1.