Nowadays, Twitter sentiment analysis has become one of the most fascinating academic disciplines. It combines methods for natural language processing with data analysis methodologies for developing such systems. This research presented an efficient approach for analysing Twitter sentiments. A Hybrid Optimization Model using Steady State Genetic Algorithm (HOM-SSGA) for Twitter Polarity Analysis is designed in this article. The suggested approach used a machine learning algorithm to identify the positive and negative polarity of tweets. During the training stage, the proposed system represented the input-labelled tweets using various approaches and feature sets. In this study, the suggested system includes three primary phases: data gathering, preparation, and sentiment categorization from precompiled tweets. The TwitterSanders-Apple 2 database was used for preliminary sentiment analysis on Twitter. On average, the raw data gathered tweets have more noise in the form of positive emojis, punctuation marks, and negative emojis, which were removed to get a more accurate result. The work employs a steady-state genetic algorithm for attribute selection and a cuckoo search for extracting features. The experimental findings and statistical assessment confirm that the suggested technique surpasses the current methods such as Artificial Neural Network (ANN), Bidirectional Encoding Representation from Transformation (BERT), Term Frequency-Inverse Document Frequency (TF-IDF), and Feed Forward Back Propagation Neural Network (FFBPNN). The suggested technique has improved effects for the analysis of social network and media data in future studies.