The generation and replication of network traffic are essential tasks for testing, analyzing, simulating, and evaluating the behavior and efficiency of systems, protocols, applications, and network services. However, it faces various challenges and limitations when generating or replicating network traffic in a realistic, responsive, and scalable manner, using appropriate models. This work proposes a modeling of a program for generating and replicating network traffic, called Forensic Training Traffic Generator (ForTT-Gen), aims to address the challenge associated with generating and replicating network traffic generated by malware by creating files that can be used in the analysis of network packets inherent to forensic analysts' training. ForTT-Gen is a tool that can generate and replicate network traffic using a hybrid model that combines the replication of real data and the generation of synthetic data through statistical techniques. Experimental results demonstrate the model's ability to reliably. Accurately reproducing the statistical patterns of the original traffic, is evidenced by a determination coefficient of 1 and a Pearson coefficient of 0.9, all within a 0,58 confidence interval with 95% certainty.