Determining the appropriate direction and location of urban development (as a necessity) is a vital factor in effective urban management. One of the most important threats to infrastructures, buildings, and the overall urban life, especially in cities facing water stress, is land subsidence (vertical displacement of the ground surface). The long-term damages of subsidence and its role in intensifying the destructive effects of high-energy and short-term events (such as earthquakes, floods, explosions, etc.) are very important and cannot be ignored, but due to its hidden and gradual nature, it is often neglected. Parameters with different weights affect subsidence, which, based on the conditions of that specific location, is unique. This study, using a quantitative-exploratory research method, aims to use machine learning and, by applying modeling based on inferential statistics and historical data with the help of Python (version 3.6), to determine the weighted impact of the factors influencing land subsidence. After identifying the influence weight of each factor on subsidence in each region, while informing urban management to eliminate or reduce the effect of those factors, with the help of the mathematical and algorithmic model it is possible to predict the future trend of each factor and the overall subsidence, so that the possibility of forecasting the future strategic development of the city based on avoiding subsidence risks will be provided.