This study examines the differences in accuracy between the hybrid convolution neural network and convolution neural network approach for recognizing platelets, white blood cells, and red blood cells (RBC) from blood samples. For the study, forty samples were collected. There are two sets with twenty samples each. Group 2 makes use of hybrid convolutional neural network technology, while Group 1 utilizes the convolution neural network technique. Once the dataset was imported in accordance with the study protocol, Google Collab software was used to produce the code for the convolutional neural network. When calculating sample size, the g power pretest power value is set at 80%, alpha 0.02. The prediction algorithms were assessed using SPSS, a statistical analysis tool. The hybrid convolution neural network technique produces an accuracy of 97.19% based on simulation data, whereas the convolution neural network method produces an accuracy of 94.98%. A difference smaller than 0.05 (p < 0.05) is indicated by the significance level of 0.0. As a result, the two groups are significantly connected. The hybrid convolutional neural network (CNN) outperforms the convolution neural network in blood sample identification when it comes to identifying platelets, white blood cells, especially red blood cells (RBC, WBC).