Agriculture, being the cornerstone of all nations, plays a pivotal role in providing sustenance and essential raw resources. Its paramount significance lies in serving as a fundamental food source for humanity. Consequently, the detection and prevention of crop diseases have emerged as a critical concern in the agricultural domain. Crop diseases, often triggered by pests, insects, and pathogens, can substantially diminish agricultural productivity on a significant scale if left uncontrolled. The detrimental impact of these diseases poses a considerable challenge for agriculturists, leading to substantial losses. Monitoring vast cultivated areas, spanning across acres, becomes a cumbersome task for cultivators due to the need for regular supervision and early detection of potential issues. Many deep learning-based methodologies are already proposed but they are not providing an up-to-mark accuracy in real-time scenario. Deep learning model performance mainly depends on the hyperparameter they are using in model training. Finding the proper hyperparameter for the model is a huge challenging and time-consuming task. In the proposed deep learning model, a genetic algorithm approach is used to obtain a proper hyperparameter. Due to the use of genetic algorithm, a metaheuristic-based optimization technique results in an increment of model accuracy from 97% to 98.5%, which is a huge achievement. The proposed system additionally notifies the farmer about crop diseases, enabling them to take necessary actions.