Abstarct: Abstract Orthoptera is one of the most important orders of insects in terms of Biodiversity and a large number of harmful species. In this order, more than 29,000 species have been identified so far, and in the Caelifera suborder, 350 species in the form of 110 genera and 6 families have been collected and identified from different areas of Iran. In this research, which was conducted from March 2019 to December 2021 in the Garmeh-Jajram region of North Khorasan province, a total of 22 species of grasshoppers of the Caelifera suborder belonging to 13 genera, 5 subfamilies, and 2 families, were collected and identified. Sphingoderus carinatus and Sphingonotus octofasciatus species from the Oedipodinae subfamily, Ramburiella turcmana species from the Gomphocerinae subfamily and Acrididae family, Dericorys annulate and Dociostaurus Hauenstein species from Dericorythinae subfamily and Dericorythidae 5 family for the locust fauna of New North Khorasan Province, as well as all the collected species, in the collection of Entomology of Zoological Museum of Prof. Jalal Afshar, is kept in Department of Botany of Tehran University of Agriculture and Natural Resources (Karaj). In addition, in the second part of this research, to estimate the spatial distribution of grasshoppers of the Caelifera suborder, an artificial neural network was used. The data related to the population density of locusts were collected and obtained from different habitats of Jajarm and Garme region in the year 1400. In this study, longitude and latitude variables were used as input variables. Also, the changes in the population of grasshoppers of the Caelifera suborder were used as output variables. The network used was an optimized multilxayer perceptron (MLP) network. To evaluate the ability of neural networks used in distribution prediction, a statistical comparison of parameters such as variance, statistical distribution and mean between spatially predicted values by the neural network and their actual values were used. One of the most important steps in creating artificial neural networks is data preprocessing. Lundberg-Marquardt training algorithm was used to implement, train and test the network using MATLAB R2016a software. To optimize the used neural network, trial and error were used and the number of hidden laxyer neurons and the type of threshold function were determined and optimized. The drawn maps showed that the distribution of grasshoppers of the Caelifera suborder is cumulative. Maps obtained from artificial neural networks can help planners to apply for pest control programs.