The utilization of machine learning in healthcare is widespread, aiding doctors and clinicians in various capacities. One of the critical implementations of machine learning in healthcare is the creation of an automated diagnosis system capable of predicting outcomes equivalent to or surpassing those of human experts. This study illustrated the optimistic outcomes using the proposed Blending with Meta Majority Voting (BwMMV) method as a cutting-edge and highly effective ensemble learning technique. Essentially, the Local Binary Pattern Histogram (LBPH), a straightforward and effective texture operator, was utilized to convert the blood smear images into their underlying texture characteristics. The proposed approach BwMMV involves creating a new dataset by using the predictions generated by the 8 base classifiers, also known as an intermediate dataset. This intermediate dataset is then used to develop five additional models using 5 distinct machine-learning algorithms, called meta-models. The final decision is made using a majority voting technique, where the outcome with the highest number of votes is selected as the final choice. The proposed technique has been found to be highly effective and resilient in terms of both performance and robustness. It has been demonstrated to be a superior approach for an automatic diagnosis system when compared to conventional hard voting and blending techniques.