This study compares the performance of Convolutional Neural Network with Naive Bayes for detecting iron deficiency through color space intensity and color characterization of the images to improve accuracy. Materials and Methods: The study employs palpable palm images from Mendeley Data as its dataset. The Conventional Neural Network and Naive Bayes is utilized to analyze and classify whether the patient is anemic or not. Sample sizes are determined using specific method, which ensures statistical power and significance. Results: The results demonstrate that Convolutional Neural Network achieved significantly better than Naive Bayes. in terms of accuracy when classifying palpable palm images for the application of iron deficiency detection using color space intensity and color characterization. The statistical analysis reveals a substantial difference p-value = 0.002; Independent Sample t-Test, p is less than 0.05 between the two Algorithms. Conclusion: This study confirms the higher accuracy of Convolutional Neural Network over Naive Bayes in effective iron deficiency detection based on color space intensity and color characterization of the palpable palm images classification.