This paper presents a 3-D search grid allocation algorithm for source localization and beamsteering. Data-driven beamsteering systems can benefit from algorithm designs that restrict the processing to a finite set of predefined direction of arrival (DoA) grid points. However, this scenario requires careful definition of the distribution of the search grid points to achieve high spatial filtering performance. The non-linear relationship that exists between time-delay of arrival (TDoA) and degree-oriented DoA means that a generic uniform grid distribution approach is not adequate for accurate localization of the source direction in the time domain. The aim of the proposed search grid allocation algorithm is to maximize the efficiency of localization by exploiting the sensitivity variation in temporal and spatial aspects. Experimental results show that the proposed search grid allocation outperforms the uniform grid allocation in terms of TDoA accuracy and steering vector estimations.