Objective: To evaluate diagnosis accuracy of immunos algorithms for classification of migraine, tension-type and cluster-type of headaches by using our website based diagnosis survey expert system. Methods: The headache diagnoses of eight-hundred and fifty patients from three different cities in Turkey were evaluated by using immunos algorithms. Data were collected through our website based diagnosis survey expert system under the guidance of neurologists. Results: Immunos algorithms for diagnosis have the maximum accuracy of 95.65% which can be used for classification. Conclusion: It is possible to classify primary headaches with immunos algorithm and using our website based diagnosis expert system will be helpful for neurologist in order to obtain precise results as well as easy information sharing.