Iran is one of the largest producers of oil and gas in the world. The use of smart manufacturing approaches can lead to better performance and less costs of the well drilling process. One of the most important issues during the drilling operation is the wellbore stability. Instability of wellbore can occur at different stages of a well's life and inflict heavy financial and time damages on companies. Selecting a proper drilling mud weight, which is a controllable factor, can prevent lots of these damages. The main goal of this research is presenting a drilling mud weight estimator for Iranian wells using Deep-learning techniques. Our Iranian dataset only contains 900 samples, but efficient deep models usually needs large amounts of data to obtain acceptable performance. Therefore, the samples of two datasets related to the United Kingdom and Norway fields are also used to extend our dataset. Our final dataset has contained more than half million samples that has been compiled from 132 wells of three fields. Our presented mud weight estimator is an artificial neural network with five hidden layers and 256 nodes in each layer that is able to estimate the mud weight for new wells and depths with the mean absolute error (MAE) of less than ±0.039 pound per gallon (ppg). In this research, the presented model has been challenged in real-world conditions. The results have shown that our model can be reliable and efficient in the real world.