News television stations are reporting current events from multiple fields such as political, social, sports and so on. Having a report that classifies the news generated by the journalists would significantly benefit the management of the news station by assisting at resource allocation, editorial content strategy, performance measurement and audience insight. This paper proposes a supervised machine learning model that can automatically classify news using text data from a Romanian branch of a multinational news network that currently does the news classification process manually. First, we collect the data in the form of Romanian-language news stories from the media asset management system of the news station. With this data, we then create and pre-process the dataset, vectorize the text data into numerical values using both CountVectorizer and TF-IDF, train the model using Multinomial Naïve Bayes algorithm and evaluate the results. The model evaluation shows good performance, the classification accuracy varied from 79 to 82%, depending on the vectorize method applied for the text feature extraction. A better accuracy was achieved when CountVectorizer method was applied. To provide additional context for the results of our news classifier application, we extended our analysis to include a Support Vector Machine (SVM) model.