Alarm management systems are an essential part of industrial units controlled by the operator using distributed control systems. Alarms are raised to inform the operator of abnormal situations, which could be due to the process variables exceeding the specified threshold or a fault. When an alarm is raised, the operator must intervene immediately and find the alarm's root cause. Otherwise, the fault could propagate through the process, causing damage to other process equipment and even shutting down the entire plant. On the other hand, due to the propagation of faults between the equipment, the operator faces many alarms in a short time, which causes the inability of the operator to handle the current situation. Alarms are the operator's guide in finding the cause of the faults. But due to the interaction and complexity of the system, there is no one-to-one relationship between alarms and faults. So we need a mechanism that can help the operator find the root cause of the alarm. Also, by predicting the next alarm and its occurrence time, the operator can manage the critical situation. In this study, the objective is to model the relationships between alarms to find their root cause and predict the next alarm with the highest possibility and the time of its occurrence. For this purpose, deep neural networks have been used, and the proposed models have been implemented on the well-known Tennessee-Eastman process. For training the neural network, the alarms are first pre-processed and organized in sequences. Since the neural network input must be numerical vectors, and the alarms are in textual sequences, the word embedding model has been used. The alarm sequences have been converted to numerical vectors. Two deep neural networks have been proposed to find the alarm's root cause with high accuracy using a small amount of data. The network training process is also fast. Then, two deep neural networks, with three inputs and three outputs, were used to predict the next alarm and its occurrence time. The sequence of alarms, the hour and minute of the alarms are the neural network input and the next alarm with the highest probability of occurrence, as well as the hour and minute of its occurrence are the outputs. Acceptable accuracy was obtained for this section as well.