Damage to the crops occurs due to insects eating the leaves, environmental changes, irregular fertilization, and improper use of pesticides decreasing yield. It is important to identify the percentage of damages in crop that will help in selecting the quantity of pesticide used to treat the damage, and in predicting the change in the yield of the crop. So far, the research efforts handled this problem by collecting datasets from near-field and far-field images of damaged crops and then training Deep Learning Model to differentiate healthy leaf and Unhealthy leaf. To the best of our knowledge, there is no deep learning model has been trained to predict and classify the level of damage in the soybean leaves. Therefore, the main aim of this thesis is to propose a deep learning model that predicts the classes of damage from scale of one to five and also identifies healthy leaf from unhealthy leaf. The proposed model analyses dataset containing non-healthy leaves and healthy leaves and estimates the performance of classification methods. This analysis allows the model to predict different damage caused by insects and natural calamities to leaves, therefore aiding the agricultural professional to corrective actions based on specific class of damage.