Around the globe, the assessment of subjective answers plays an important role in the education system. This assessment helps to evaluate the performance of a student to a certain extent. However, there are a few shortcomings to the conventional human evaluation method, which makes it less practical. The manual evaluation, though essential, is a bit of a tiresome and time-consuming task. Currently we have a reliable system for automatically assessing objective answers. Over the years, researchers have put their efforts into the automatic assessment of descriptive answers. But this research is still at its infancy. In this research work, we demonstrate the use of natural language processing and machine learning techniques for the automatic subjective assessment. We demonstrate and compare semantic similarity measures. The available pretrained deep learning models, like sentence transformers, are used in this research. Finally, we have developed regression models and demonstrated the performance of the proposed approach.