Abstarct: Abstract This dissertation analyzes and predicts the changes and future trends of the deltaic formations at the mouths of three key tidal creeks along the Makran coast of Iran—Tang, Galak, and Meydani. These creeks, which are simultaneously influenced by wave and tidal forces, are connected to the ocean through estuarine inlets and, due to this interaction, demonstrate high sensitivity to physical environmental changes, such as sediment delta growth or erosion. Identifying patterns of deltaic transformation in such wave–tidal systems is of critical importance for planning sustainable regional coastal management. To achieve these objectives, hydrodynamic, meteorological, geomorphological, and atmospheric parameters were collected over a 20-year period (1999–2019) from reputable international databaxses and local monitoring stations. Using three well-known machine learning algorithms—Decision Tree (DT), Multi-laxyer Perceptron (MLP), and Support Vector Machine (SVM)—these datasets were analyzed and modeled. This modeling led to the identification of 34 influential variables contributing to deltaic changes, encompassing a wide range of environmental factors. Findings indicate that among these 34 variables, sea level rise, barystatic pressure, precipitation, and the frequency of rainy and stormy days have the most significant impact on delta variation. Moreover, trend analysis shows a steady decline in the volume of deltaic deposits at the studied creek mouths over the past two decades. Among the applied models, the Decision Tree algorithm yielded the highest predictive accuracy with an error margin of less than 5%. The results further demonstrate that physical characteristics of the creeks significantly influence model performance. Specifically, larger inlet widths and greater creek surface areas correspond to improved model accuracy. Conversely, the presence of structural features like the tombolo at Tang creek and the narrow inlet of Galak creek disrupt sedimentation patterns and reduce the accuracy of the predictive algorithms.