The gradient-dependent correlation scale method was used to construct the temperature vertical structure of yellowfin tuna fishing ground based on Argo real-time observation profile. The results showed that, compared with the traditional method, the root mean square error of the vertical temperature profile constructed is significantly reduced. At the depth of the thermocline, the maximum difference between the two was 0.45℃. Besides the thermocline, the temperature deviation between the constructed temperature profile and the actual observed true value was great (about±0.5℃), the other water layers were almost zero. The results of statistical analysis in August 2017 showed that the suitable temperature range of yellowfin tuna in the near surface layer (5 m) of the western and central Pacific Ocean is 28.5-29.5℃, and with the increase of depth, the temperature gradually decreases, and the suitable temperature range is about 11.0-12.0℃ at 300 m. There was obvious thermocline in each fishing point, the corresponding average thermocline upper and lower boundary depth and thermocline strength are about 80 m, 270 m and 0.08℃/m respectively, and the thermocline strength is approximately proportional to the catch. These conclusions further verify the reliability of the method.