Nuts are nutritionally valuable for a healthy diet but can be prone to rancidity due to their high unsaturated fat content. Nutrient content of nuts is an important component of their health benefits but measuring both rancidity and nutrient content of nuts is laborious, tedious and expensive. Hyperspectral imaging has been used to predict chemical composition of plant parts. This technique has the potential to rapidly predict chemical composition of nuts, including rancidity. Hence, this study explored to what extent hyperspectral imaging (400–1000 nm) could predict chemical components of Canarium indicum nuts. Partial least squares regression (PLSR) models were developed to predict kernel rancidity using peroxide value (PV) for two different batches of kernels, and macro- and micronutrients of kernels using the spectra of the samples obtained from hyperspectral images. The models provided acceptable prediction abilities with strong coefficients of determination (R2) and ratios of prediction to deviation (RPD) of the test set for PV, first batch (R2 = 0.72; RPD = 1.66), PV, second batch (R2 = 0.81; RPD = 2.30), total nitrogen (R2 = 0.80; RPD = 1.58), iron (R2 = 0.75; RPD = 1.46), potassium (R2 = 0.51; RPD = 0.94), magnesium (R2 = 0.81; RPD = 2.04), manganese (R2 = 0.71; RPD = 1.84), sulphur (R2 = 0.76; RPD = 1.84) and zinc (R2 = 0.62; RPD = 1.37) using selected wavelengths. This study indicated that visible-near infrared (VNIR) hyperspectral imaging has the potential to be used for prediction of chemical components of C. indicum nuts without the need for destructive analysis. This technique has potential to be used to predict chemical components in other nuts.