MOST OF THE CLASSIFICATION ALGORITHMS HAVE BEEN DEVISED TO CLASSIFY LONG TEXTS, SUCH AS EMAIL AND WEB PAGES WHICH OVERSHADOWED THEIR EFFECTIVENESS ON SHORT AND SOMETIMES INFORMAL TEXTS. IN THIS PAPER, WE EVALUATED THE ACCURACY OF FOUR MAJOR CLASSIFICATION ALGORITHMS ON PERSIAN SHORT TEXTS. THESE ALGORITHMS ARE NAÏVE BAYES, K-NEAREST NEIGHBORS, DECISION TREES AND SUPPORT VECTOR MACHINE. FIRST, WE BRIEFLY INTRODUCE THEIR OVERALL METHOD AND PROVIDE SOME BASIC INFORMATION, AND THEN, WE APPLY THESE ALGORITHMS TO ONE SPECIFIC DATASET TO MEASURE THEIR EFFECTIVENESS. RESULTS SHOW THAT THE NAÏVE BAYES ALGORITHM FUNCTION COMPARATIVELY BETTER THAN THE OTHERS, WHILE KNN ALGORITHM HAS THE LEAST ACCURACY.