Traditional methods for assessing student performance in maritime simulator training usually rely on instructor observations or post-manoeuvre de-briefing. This study explores the potential of applying image recognition technology to assess student performance during critical manoeuvres such as Williamson turn in a nautical simulator training. By utilizing image recognition through convolutional neural network algorithm, a system is proposed that can analyze key manoeuvre aspects using visual data from the simulator, thus aiding students in assessing their performance. Analysing these visual cues with an image recognition algorithm could potentially serve as a component of a learning analytics dashboard (LAD) for maritime simulator training.