With the increase in population and the growth of demand for agricultural products, it is necessary to use new technologies to increase productivity. Wireless sensor networks, as one of the key technologies in smart agriculture, enable continuous and accurate monitoring of environmental parameters such as temperature and humidity. These systems provide farmers with the possibility of optimal management of resources by wirelessly transmitting information to control centers. This technology faces important challenges, such as information security and efficiency in heterogeneous and distributed environments. Reviewing the previous research found that the existing articles have not comprehensively addressed the security and efficiency of these systems. Therefore, this paper aims to provide a comprehensive solution to improve the efficiency of wireless sensor networks in smart agriculture. Also, appropriate hardware for accurate monitoring of environmental parameters, including temperature and humidity sensors, has been introduced and implemented. To improve the accuracy of data processing and control, the optimized convolutional neural network with a genetic algorithm was used for fault detection, and the proposed control algorithm was used to control the speed of agricultural motors. The proposed security protocol was formally and informally evaluated using the Scyther tool. The results showed a significant improvement in information security and reduced energy consumption. Also, by using vibration data, we achieved 99.92% accuracy for test data, and by simultaneously using vibration and acoustic data, we achieved 99.96% test accuracy, indicating the data combination's effectiveness for fault detection.