Eggs are considered an important food item, containing protein as the main nutrient. The aim of this study was to non-destructively detect protein content in eggs by visible (VIS)/near-infrared (NIR) reflectance spectroscopy. VIS/NIR raw reflectance spectra of fresh egg samples were acquired in the wavelength range of 400 to 1000 nm. The raw spectrum was pretreated with multiplicative scatter correction (MSC) and first-derivative (1-D) methods and step-wise regression discrimination method was used to select the optimal wavelength combination to establish multi-linear regression (MLR) models. Full cross-validation was used to validate the model. The results showed that after MSC treatment of VIS/NIR reflectance spectra, the MLR model, based on ten optimal wavelengths (400, 403.16, 407.9, 714.6, 715, 715.58, 970.4, 970.75, 973, and 974.45 nm) produced optimum calibration and validation results. For the calibration result, the correlation coefficient (R) was 0.92 and the standard error of calibration (SEC) was 0.47%. The model provided good prediction values for egg protein content with the correlation coefficient of cross validation (Rcv) at 0.89 and standard error of cross validation (SECV) at 0.47%. This study demonstrated that the VIS/NIR reflectance spectral technique provides a good prediction of the protein content in fresh eggs, and the VIS/NIR technique has potential applications in rapid detection of egg nutrients.