Abstract: The present study investigated the quality changes in fresh macadamia nuts during short-term storage by examining the effects of various storage temperatures and humidity levels (30 ℃-RH80%, 35 ℃-RH80%, 40 ℃-RH80%, 30 ℃-RH90%, 35 ℃-RH90%, and 40 ℃-RH90%) on the moisture content of the pericarps, nut-in-shells, and kernels. In addition, the cracking and mold rates of pericarps, acid value, peroxide value, iodine value, total phenol content, and total sugar content were evaluated. A quality prediction model for the short-term storage of fresh macadamia nuts was developed using a backpropagation (BP) neural network, and its predictive capability was evaluated using a test set. The results showed that water loss was the fastest in the 35 ℃-RH80% storage group. The cracking rate of the pericarps stored at 35 ℃ increased the most and was significantly higher than that of the nuts treated at 30 and 40 ℃ (P<0.05). The mold rate of the pericarps stored at 30 ℃ increased the most and was significantly higher than that of nuts treated at 35 and 40 ℃ (P<0.05). Both the acid and peroxide values increased during storage; at the end of storage, the highest acid value was 15.57 mg/100 g in the 35 ℃-RH90% group, and the highest peroxide value was 36.44 μg/g in the 30 ℃-RH80% group. During storage, the iodine value and total phenol contents first increased and then decreased. The maximum increase in the iodine value was 119.26 mg/g in the 35 ℃-RH90% storage group. At the end of storage, the iodine value in the 40 ℃-RH80% storage group was 675.72 mg/g, which was lower than that under the other storage conditions. The total phenol content at the end of storage in the 35 ℃-RH80% and 40 ℃-RH90% groups was 0.88 mg/g, which was significantly lower than that under the other storage conditions (P<0.05). The total sugar contents of the samples decreased gradually, that of 35 ℃-RH80% storage group at the end of storage was significantly lower than the other storage conditions (P<0.05). The correlation analysis showed that the input and output layers of the prediction model exhibited good correlation. The quality prediction model of fresh macadamia nuts featured seven hidden layer nodes, with correlation coefficients for acid value, peroxide value, iodine value, total phenol content, and total sugar content of 0.97952, 0.98815, 0.94869, 0.94882, and 0.97109 in the training set, respectively, indicating good prediction accuracy. Therefore, this neural network prediction model can be used to predict the quality changes of fresh macadamia nuts during postharvest transportation and storage. This study lays a foundation for the application of the neural network prediction model in macadamia nut quality prediction.