The invention discloses a vertical well flow pattern recognition method based on conducting probe array voting and belongs to the field of multiphase flow detection. The method includes: measuring voltage response signal of each probe of a conducting probe array; extracting characteristic quantity from the voltage response signal of each probe through statistical analysis and wavelet analysis; performing Z-score normalization on the extracted characteristic quality, and adopting principal component analysis (PCA) technology to extract principal component to form PCA characteristic quantity; utilizing a support vector classification (SVC) method to respectively build classification models from the PCA characteristic quantity of the voltage response signal of each probe to flow patterns of oil-water phase flow; adopting a particle swarm optimization (PSO) algorithm to optimize parameters of the SVC models; performing decision-making-level information integration based on voting on flow patterns recognized by the probes. By the method, input variable dimension is lowered substantially; compared with vertical well flow pattern recognition methods based on single probes, the method has the advantages that robustness and reliability of logging are improved, and flow pattern recognition rate is increased.