With the introduction of 5G technology, operators can now design systems that are scalable, effective and flexible thanks to previously unheard-of possibilities for virtualized and intent-based network management. Intent-based networking makes network administration easier while improving dependability and flexibility by concentrating on “what” results are sought rather than “how” they are attained. This study investigates intent prediction in 5G networks using Auto-Regressive Moving Average (ARMA) modeling. ARMA provides a linear, resource-efficient method for predicting important intent variables like bandwidth, CPU and memory needs by utilizing previous data. In situations with little datasets, ARMA’s ease of use and effectiveness are demonstrated through comparison with machine learning methods. By adding external factors and non-stationary data trends, extensions such as ARMAX and ARIMA improve predictions even further. These findings illustrate ARMA’s potential for proactive intent management, assuring optimal resource allocation in dynamic 5G contexts, and laying the framework for sophisticated machine learning techniques integration.