Drilling is an important and expensive procedure in various stages of mineral exploration projects, so that continuation of mining activities mostly depends on drilling results. Designing the optimal exploratory boreholes is very important from the perspective of budget and information value. Exploratory drilling can be designed in such a way to minimize the cost and maximize the obtained information about the concealed ore bodies. In general, drilling is done as initial drilling at the end of the preliminary exploration stage, and systematic and complementary drilling during the detailed exploration stage. The surface exploration information such as geological maps (including host rocks, alterations and faults), surface geochemical data, geophysical data, and ground observations are used in designing and locating the primary boreholes. The purpose of primary drilling is to examine the presence or absence of subsurface mineralization and to determine the approximate zones of ??the ore body. Therefore, designing optimum primary boreholes should be done in such a way to intersect the concealed ore with the highest probability. The accurate interpretation and proper integration of the surface exploration data can lead to favorable results in selecting the optimum boreholes. In past, data integration methods such as fuzzy logic, overlay index and multi-criteria decision-making methods have been used to design optimum boreholes. In this research, semi-supervised When the drilling results of primary boreholes are positive, then the complementary boreholes should be designed to achieve maximum information about the concealed ore body with the possible lowest cost. Therefore, selection of optimum locations for boreholes is a multi-criteria decision-making problem. The spatial and directional parameters of additional boreholes can be determined in two continuous and discontinuous spaces. Hence, in this research, two general approaches for designing the complementary boreholes in a continuous and discontinuous spaces were introduced. In a discontinuous mode, a set of directional candidate boreholes were designed and the score of each borehole was calculated for different criteria. Then, using a multi-criteria decision-making method, candidate boreholes were ranked, and the best one was selected as the optimum additional borehole. In this case, the optimum additional boreholes were selected sequentially and the second borehole was dependent on the result of first borehole. In continuous mode, a method was introduced on the basis of the integration of a multi-criteria decision-making method and a meta-heuristic optimization algorithm. In the later method, the decision function of the multi-criteria decision-making algorithm was considered as the objective function, and particle swarm optimization (PSO) algorithm was used to optimize the objective function. In this method, there was no need to design a set of candidate boreholes and also it was possible to design a group of additional boreholes simultaneously. Several criteria for designing the exploration boreholes can be considered, depending on the type of ore deposit, existing data, budget and the exploration stage. In this research, two proposed methods were utilized to design complementary boreholes at the Dalli Cu-Au porphyry deposit. Nine criteria were defined with various weights to design the optimum boreholes. These criteria were Cu-Au grades, Kriging estimation variance, thickness of ore body, gangue thickness, average magnetic susceptibility, drilling costs, depth of ore deposit and type of alteration hosting mineralization. Both proposed methods were used to locate four additional boreholes. The proposed complementary boreholes by these methods are located in the northwest zone of the south Dalli Cu-Au deposit. In this research by considering several exploration criteria and using various semi-supervised classification algorithm in 2D and 3D environment individual and a combination bore holes were designed. This automatic approach is important in various stages of drilling porphyry Cu-Au deposit or by changing the exploration criteria in other types of ore deposits. It is proposed to use this approach in a new software or as an individual module in the modelling and reserve estimation software.