Abstract: A study on phylogenetic relationships in euglenoids is made using the principles of cladistic analysis．In the analysis，33 genera(i．e．OTUs) of euglenoids were selected and 35 informative characters were used．Three criteria，“outgroup comparison”，“commonality principle”and“principle of character correlation”，were used in the polarity determination of character states． The sister group of euglenoids ought to be Kinetoplastids in protozoan on the basis of the flagellate structure and molecular biological evidences．Because the euglenoid fossil evidence is inadequate and such fossils，even present，have a very simple ontogenetic process，“fossil evidence”and“recapitulation law”are not used to determine primitiveness． A data matrix was constructed by the 33 OTUs and 35 characters．The data matrix was analysed with“evolutionary extremal aggregation method”．A reasonable parsimonious cladogram(lenth=251，parsimonious coefficient=0．2159，was constructed and used as a basis for discuussing the phylogenetics of euglenoids． In the cladogram，the euglenoids may be divided into five large groups．The cladistic points 5l and 48 are the most primitive euglenoids，which are phagotrophic，colourless euglenoids and in closely related to Kinetoplastids in protozoan；the point 62 is green euglenoids，which probably evolved by endosymbiosis between phagotrophic colourless euglenoids and green algae;the point 58 is the most complex group，containing advanced green euglenoids in evolution and saprophytic colourless euglenoids which probably evolved from green euglenoids by losing their chloroplasts；the point 37 is a rare group，parasitic euglenoids with many flagella． According to the phylogeny of euglenoids in the cladogram，a classification which includes one division，one class and five orders may be proposed．It is obviously different from the former systems set up by other biologists．Five subgenera of Euglena may be identified as independent genera on the basis of their characters and evolutionary lines in the cladogram.

Key words: Euglenoids, cladistic analysis