The current research questions that are raising concern security in cloud computing. Cloud storage is replacing traditional data storage in numerous companies because it provides instant knowledge of data from any location. However, the main problem stopping businesses from utilizing cloud computing is data security. Cloud computing has been one of the most talked about subjects in technology lately due to its capacity to manage a large number of resources and data. One of the biggest obstacles is scheduling tasks, and inefficient management leads to performance decline. This paper presents a unique method for productive task scheduling with improved security in a cloud computing environment. It is suggested to design the operations using a unique Convolutional neural network optimized modified butterfly optimization (CNN-MBO) approach in order to improve efficiency and shorten construction time. Second, the data is encrypted using a hybridized RSA method, ensuring safe data transfer. The performance of our suggested technique is then evaluated by simulating it using a cloudlet simulator and analyzing the evaluation findings. The suggested method is also contrasted with alternative task scheduling-based methods for a number of performance criteria, including resource usage, reaction time, and energy consumption.