For　energy-saving　and　improving　ride　comfort,　this　paper　presents　a　novel　development　of　the　longitudinal　motion　control　of　intelligent　vehicles　based　on　two　hierarchies　optimal　methods.　The　upper　method　is　the　Radau　pseudo-spectral　method　(RPM),　the　lower　method　is　the　model　predictive　control　(MPC).　The　RPM　is　used　for　energy　consumption　optimization　algorithm.　The　MPC　is　used　for　the　longitudinal　motion　control　method.　The　longitudinal　motion　model　and　energy　consumption　model　are　developed.　Based　on　the　preceding　models,　the　optimal　control　problem　of　energy　consumption　optimization　is　established,　in　combination　with　the　boundary　conditions　and　path　constraints.　Using　the　RPM　to　solve　the　problem,　the　optimal　vehicle　speed　trajectory　is　obtainded　as　desired　input　The　longitudinal　motion　control　is　completed　based　on　MPC.　Simulation　results　show　that,　in　the　case　of　a　pure　electric　vehicle　and　an　actual　planning　path,　automatic　driving　at　optimal　speed　consumes　less　power　energy　than　automatic　running　at　constant　speed,　and　verify　the　effectiveness　of　the　strategy　in　the　paper.　©　2015　IEEE.