The abnormal egg, blood-egg, has the internal optically active anomalies blood spots. A detection system based on Vis/NIR spectroscopy was designed for eggs’ internal quality inspection. Then, spectra data of 40 brown normal eggs and 60 artificial abnormal eggs obtained by injecting a small amount of blood ingredient were collected. Three analytical methods: the traditional discrimination method using blood value, discriminant partial least squares analysis (DPLS) and least squares support vector machine (LS-SVM) combined with spectral information and eggshell color information were compared for discrimination analysis. The final results showed that the discrimination rate based on LS-SVM method was 90.0% for normal eggs and that was 91.7% for blood-spot eggs, better than the results using traditional discrimination analysis methods. The results suggested that LS-SVM was an effective analytical tool in detecting blood-spot eggs with brown eggshell.