The complexity of legal texts, characterized by language and intricate reasoning, poses unique challenges for summarization. The objective of this study is to integrate abstractive summarization techniques to generate comprehensive judgment summaries. Various summarization models, including both extractive and abstractive approaches, are evaluated and compared based on their performance in distilling key information from legal documents. This paper presents an organized method for summarizing legal judgment-related case documents. Data collection is the first step where case documents are collected from the Indian Kanoon website and preprocessed to extract the text from the case document. Judgment-related keywords are identified to enhance the relevance of data. Lastly, the text is parsed and passed through two extractive summarizers (Latent Semantic Analyzer and Reduction Summarizer) to generate the initial summary and the initial summary is re-defined by a fine-tuned abstractive summarizer. The study explores the effectiveness of different summarization models in summarizing judgments from trial courts, high courts, and the Supreme Court of India. Comparative analysis of the summarization models such as facebook/bart-large-cnn, google/pegasus-cnn_dailymail, pszemraj/led-large-book-summary, xlnet-base-cased and gpt2-medium --and Fine-tuned modified Pszemraj/Pegasus-x-large-book-summary reveals varying strengths and limitations, highlighting the importance of fine-tuning and optimizing summarization techniques for accurate and comprehensive extraction of legal information. Through the rigorous evaluation the paper finds the best fit model as fine-tuned Pszemraj/Pegasus-x-large-book-summary as the final summarization model for the desired judgment summary which gives a precision of 0.9600, recall of 0.8400 and F1-Score of 0.8700.