Abstract: To investigate the quantitative relationship between the instability patterns of acoustically levitated droplets and acoustic control parameters, systematic experiments on droplet instability were conducted in a single-axis standing-wave acoustic field. Phase diagrams were constructed using the Weber number (We), Ohnesorge number (Oh), and droplet volume (V) as characteristic parameters. By varying the physical properties and initial volumes of the liquids, droplet atomization and deformation evolution under different conditions were observed, yielding three typical instability patterns: edge splashing atomization, central fountain atomization, and bubble formation. Based on the experimental data, a three-parameter statistical discrimination model—We–Oh–V—was established; training and test sets were partitioned using stratified sampling for performance evaluation. Results demonstrate that the different instability patterns exhibit distinct zoning characteristics in the parameter space. The constructed multiclass logistic regression model achieves an overall classification accuracy of 82.2% on the test set, demonstrating reliable pattern prediction capability. This study provides experimental evidence for identifying instability behaviors of acoustically levitated droplets and for regulating atomization in acoustic fields.