Over the last few years, Automated Essay Scoring (AES) technology systems have seen increasing focus. Researchers focus on AES to reduce the load of educators and provide consistency along with equity in grading essays. This paper designs and develops an AES system that makes use of NLP techniques and ML algorithms for assessing student essays. It evaluates language features-including grammar, coherence, structure, and content relevance-for scores that approximate human judgments as closely as possible. In this paper, we evaluate several scoring models, both traditional statistical and deep learning approaches, to select the single model that performs best on essay grading tasks. Another aspect is the estimation of the bias present in automated grading and how it might be ensured that fairness is guaranteed at the same time as transparency on the scoring procedure. The proposed AES system shows very promising results with high accuracy compared to the human evaluators and has the potential to bring revolution into the landscape of educational assessment by offering quick, reliable, and scalable solutions in grading.