Insect detection in the soil is important in entomology research, pest control, and insect-resource utilization, and it is more important for farmers in farming to identify the soil insects, to make the soil more fertilizer, etc. Identification of the soil insect by the naked eye or manually can be hard and time-consuming because Insects are the most diverse group of organisms, spread over half the population of all known organisms on the Earth, and only trained professionals are able to identify insects. The aim of this paper is to create an Android application for basic soil insect detection using the Inception-v4 Model, a convolutional neural network. The Android app, named Soil Insect Detection, the main design of this application is to provide detection information offline to the users. Some insects examined in this project are Ants, Termites, Grubs, and Slug. The model was tested in different orientations, image capturing angles, and backgrounds. The model gave 90% accuracy in detecting of insects in the soil and told the insect name or medicine name that helps prevent insects. Due to this, Soil Insect Detection can be a useful and portable digital application that provides basic soil insect detection and gives the benefits of this application to different fields of research and to the farmers.