Conditions that more readily convert into retinal diseases such as Choroidal Neovascularization (CNV), DRUSEN, and Diabetic Macular Edema (DME) concurrently cause fluid accumulation in the retina, thereby creating a significant threat to vision and quality of life. Early detection is paramount to prevent permanent damage. Optical Coherence Tomography (OCT) provides non-invasive high-resolution imaging of the retina in cross-section and strengthens the power of diagnosis. However, the evaluation of OCT images is a time-consuming challenge and varies with the level of clinical experience. We trained three convolutional neural network (CNN) architectures, VGG16, InceptionV3, and an ensemble of both CNNs to identify retinal fluid and proposed a new automated technique for retinal fluid detection based on deep learning. The results of the ensemble model were the best, yielding 97.34% accuracy, 97% precision and 97% recall which outperformed the individual models. This shows that the ensemble model performs well for fluid detection in various retinal diseases.