The investigation of the coordination of multi-finger synchronous key-touching in professional pianists holds significant importance in evaluating the training effectiveness for piano novices, human-machine collaboration control of piano teaching robotic hands, among other domains. Currently, the analysis of characteristics in piano key-touching actions primarily utilizes empirical formulas, focusing on single-finger key-touching movements. Addressing the limitation of the empirical formula's applicability and the lack of accurate and practical modeling methods for multi-finger synchronous key-touching actions, a modeling approach based on the sparrow search algorithm (SSA) and back propagation (BP) neural network is proposed. Given the uniqueness of the weak-finger, the research concentrates on the key-touching actions of the weak-finger assisted by the middle finger. Data were collected on multi-finger synchronous key-touching in weak-finger piano playing from twelve professional pianists utilizing a Leap Motion sensor. A coordination model, based on a BP neural network for multi-finger synchronous key-touching in weak-finger piano playing, was established and optimized using the genetic algorithm (GA) and SSA. The results demonstrate that the SSA-BP neural network-based coordination model exhibits enhanced predictive accuracy, with a root mean square error (RMSE) of 4.7226°. Based on this model, the key-touching movements of the weak-finger MCP joint can be accurately predicted from the key-touching movements of the middle finger, providing an evaluative method and scientific guidance in training for multi-finger synchronous key-touching in weak-finger piano playing among learners.