As Bangladesh is fully dependent on agriculture so identifying the diseases of the plants will aid the farmers for better productivity. This paper elucidates the recognition of rose leaf disease which is based on deep learning algorithms. By using deep learning algorithms, the defective images of leaves can be identified precisely which may further help the farmer to identify the leaf disease accurately, and then they will immediately take precautions according to the disease. In this paper, first, we have collected the images from the nursery and labeled each disease after preprocessing them in such a way that it is fitted to the particular model that we used. Later, we have modified and implemented the CNN model compatible with our dataset to get a better outcome. By doing so, we can identify the diseased leaves 96% accurately.