The quantitative intersection technique is widely used in oil and gas exploration and also plays an important role in the recognition of reservoir fluid. This paper deals with the multi-parameter optimization method. Through calculation, highly-sensitive parameters can be obtained and, on such a basis, fluid attribute sensitivity is analyzed, thus avoiding multi-interpretation and uncertainty in multi-parameter choice. With the global survey-based 25-data set published by Castagna and Smith as the model, the authors made 3D intersection analysis with optimized parameters, and attained good result in reservoir recognition.