The aim of geophysical inverse problem is to determine the

spatial distribution and depths to buried targets at a variety of scales;

it ranges from few centimetres to many kilometres. To identify ore bodies,

extension of archaeological targets, old mines, unexploded ordnance

(UXO) and oil traps, the linear geomagnetic inverse problem resulted

from the Fredholm integral equation of the first kind is solved using

many strategies. The solution is usually affected by the condition of

the kernel matrix of the linear system and the noise level in the data

collected. In this paper, regularized Kaczmarz method is used to get a

regularized solution. This solution is taken as an initial solution to bat

swarm algorithm (BA) as a global swarm-based optimizer to refine the

quality and reach a plausible model. To test efficiency, the proposed hybrid

method is applied to different synthetic examples of different noise

levels and different dimensions and proved an advance over using the

Kaczmarz method.