The soluble solid contents (SSC) and ratio of sugar and acid are main indices for evaluating the quality of apples. Comparison between diffuse mode and transmittance mode was explored for determination of SSC and ratio of sugar and acid using online detection equipment. 180 samples of watercore and Fuji apples were divided into the calibration and prediction sets (136∶44) for developing calibration models and assessing the performance of the models. The visible and near infrared (NIR) spectra of watercore and Fuji apples were collected at the speed of 5 samples per second. The partial least square regression models were developed with the spectra preprocessed with multiplicative scatter correction (MSC), standard normal variate transformation (SNV), and baseline correction (BC). 44 new samples were used to evaluate the predictive abilities of the models. By comparison, the transmittance mode was superior to diffuse model. Because the transmittance mode could restrain stray light more effectively. The prediction correlation coefficients (Rp) of the model of soluble solid contents (SSC) was 0.936 and prediction (RMSEP) of the root mean square errors was 0.476°Brix；The prediction correlation coefficients (Rp) of the model of ratio of sugar and acid was 0.785 and prediction (RMSEP) of the root mean square errors was 10.94；The results showed that online detection of soluble solid contents and ratio of sugar and acid of apples based on Vis/NIR diffuse transmit-tance spectroscopy was feasible. The research can provide technical support and reference basis for grading of mass fruits quality.