The piano concerto “Yellow River”is one of the most influential works in Chinese piano concertos. This paper designs an ANN beat classification model based on audio analysis, calculates the beat cycle of the piano piece, and synthesizes the beats of the piano piece using the PSOLA algorithm, presenting an audio analysis of the piano concerto “The Yellow River”. Collate the compositional techniques and layout elements, musical characteristics, and points of nationalistic performance styles of the piano concerto“Yellow River”. Verify the universality of the audio analysis technique for the piano piece by identifying beats from different musical styles. The validity of the ANN-based audio analysis is determined by the piano keys and the audio selections of the piano concerto of “The Yellow River”. Select the classic performance version of the “Yellow River” piano concerto and count the total duration of different performance versions. Combining the designed audio analysis technique for piano compositions, the performance speeds and average speeds of each version are visually organized. Analyze the creative expression of the performance speed on the emotion and style of the piano concerto of “The Yellow River”. In the whole work, the solo piano melody has 187 beats in total. Yin Chengzong, Lang Lang, and Wan Jieni have the same average speed as 70.3 beats, 72.4 beats, and 59.3 beats, respectively. The overall tempo design of Yin Chengzong and Lang Lang has obvious peaks, and Yin Chengzong’s playing speed has a higher degree of ups and downs. The music has a more fluid feel, and the playing speed is used to demonstrate the immense momentum.