The　consensus　tracking　control　of　a　class　of　nonlinear　underactuated　multi-agent　systems　with　uncertainties　is　studied　in　this　paper.　A　radial　basis　function　neural　network　(RBFNN)-based　direct　adaptive　control　algorithm　is　designed.　Unlike　many　previous　articles　in　which　a　neural　network　is　employed　to　identify　the　unknown　nonlinear　functions　in　the　system　model　or　controller,　this　method　directly　approximates　the　ideal　control　law　by　a　neural　network,　making　the　control　law　simpler.　Based　on　the　neural　network　direct　control　algorithm,　a　controller　with　a　single-parameter　learning　scheme　is　designed.　This　new　control　method　reduces　the　computational　burden　by　reducing　the　amount　of　computational　data.　Finally,　the　closed-loop　system　is　proved　to　be　ultimately　uniformly　bounded　stable;　the　application　examples　of　the　controllers　are　given　by　simulation.