A potential Seebeck microprobe apparatus is described such that a profile of Seebeck coefficients can be detected on a material sample surface for thermopower investigations. Due to its spatially resolved limit on detecting small inhomogeneities of dopants or composition changes, we here propose a constructive combination of numerical modeling and practical measurement to improve the spatial resolution by deconvolution algorithm. The relevant transfer function, obtained from numerical calculations, was successfully applied on real measurement data. Besides, an improvement in detecting the 1.5 microm inhomogeneity has been achieved by applying one transfer function of the known tip setup on a set of Seebeck line scan signals, obtained with 6 microm tip-sample contact length, 3 micros signal capture time after contact, and 0.75 microm scan period. The result showed that a system theoretical approach in terms of deconvolution algorithm could theoretically enhance the spatial resolution anyway, but is practically limited by the systemic preconditions (tip size, signal capture time, and scan period) and the signal to noise ratio of the captured thermovoltages. Therefore, we provide a clear understanding of systemic preconditions and their impacts on the detection of small inhomogeneities and thus the performance in the thermopower analysis.