Recent advancements in opinion mining on X (formerly Twitter) have focused on analyzing tweets to determine user sentiments about specific events. Many researchers have adopted machine learning and deep learning methods for this task. This study proposes a new approach that combines the C4.5 algorithm, fuzzy rule patterns, and convolutional neural networks (CNNs) for sentiment analysis. The approach involves six steps: first, preprocessing the data to remove noise; second, vectorizing the tweets using word embeddings; third, using CNNs to extract key sentiment and contextual features; fourth, fuzzifying the CNN outputs with a Gaussian fuzzifier to handle ambiguous data; fifth, applying a fuzzy version of the C4.5 algorithm to generate a fuzzy decision tree and rule base; and finally, classifying new tweets using fuzzy General Reasoning based on the rule base. This hybrid method effectively handles uncertainty and imprecision in tweet data by combining CNN and C4.5 technologies with fuzzy logic. The fuzzy rule system consists of three phases: fuzzification, fuzzy C4.5-based inference, and defuzzification via General Reasoning. The approach was tested on a Hadoop-based cluster of five nodes, allowing it to manage large-scale data. Experimental results showed that the proposed model outperformed other classification methods on the COVID-19_Sentiments dataset, with notable improvements in precision (94.56%), and an overall classification accuracy of 95.15%. This demonstrates its effectiveness in handling large datasets and improving classification performance.