In order to improve the ill-condition of normal matrix and achieve ambiguity fast resolution, an inertial navigation system (INS) aided GPS integer ambiguity algorithm is proposed. INS single-difference pseudo range, GPS double-difference pseudo range and GPS double-difference carrier phase is used to get the float ambiguities and their variance-covariance matrix. Then LAMBDA algorithm is used to get the integer ambiguity. The reason why INS aided information can improve collinear and ill-condition is analyzed. Finally, an experiment is conducted with the GPS test data and the inertial navigation data from simulation. The results show that, with INS aided information, the condition number of normal matrix is dropping two orders of magnitude, 55.81% reduction of ADOP and 75.44% less time to resolve integer ambiguity. The method is feasible.