Machine learning possesses a wide variety of applications in this modern era. One of the most important of them is smart agriculture. Smart agriculture mainly made use of several advanced technologies. The main purpose of smart agriculture is to reduce the amount spent on agriculture and thereby reduce human effort without compromising the output. Constantly changing climate and surge in population acts as a challenge in today’s life. It creates a negative impact on food security also. It is here that smart agriculture plays an important role. Even though several types of prediction systems occur in agriculture, performance of the system should be taken care of. Hence, there is the need for an efficient prediction system. This paper deals with an analysis on prediction systems with different algorithms, out of which the proposed method using GRU with Adam optimization performs well in terms of accuracy, precision, recall, R squared score, mean absolute error, and root mean squared error whose values are 1.0,0.9826,0.9818,0.9527,0.1568 and 1.2617, respectively. The result obtained is compared with other algorithms such as “simple Recurrent Neural Network”, “Artificial Neural Network”, “Long Short-Term Memory”, “Logistic Regression”, “Convolutional neural network-Bi directional Gated Recurrent Unit”, “GRU-LSTM”, and “CNN-GRU”.