For lack of gradient tensor large field error in magnetic localization system, a firstorder magnetic gradient tensor localization method is proposed based on immune cloning algorithm. First, the crossform sensor model is utilized for deriving the firstorder magnetic gradient tensor localization principle. Then, an error gradient matrix calculating method is proposed by using the magnetic field decay law and the immune cloning algorithm is optimized. To verify the effectiveness of the new algorithm, three different routes are used in Matlab for onestep metric tensor localization and optimization algorithm localization calculation and to verify the new localization algorithm through actual localization experiments .It is found that the firstorder magnetic gradient tensor localization method based on the immune cloning algorithm can effectively improve the accuracy of the magnetic positioning system.