Abstract—Diabetes is a type of metabolic disease that causes a high blood glucose level that wildly found in many countries. Blood glucose measurement is necessary for diabetes patients to check how much glucose is present in the blood. The typical method to measure blood glucose level is an invasive method that gives a highly accurate result, but the patients get suffer from physical pains and it has a higher risk of infection. This research presents an alternative method, which is noninvasive blood glucose monitoring by means of Near infrared sensors based on 940nm near infrared spectrum and an artificial neural network analysis. The concept is focusing on glucose absorbance detection when the spectrum passes through the patient’s finger. In processing the signals, the wavelet’s transformation is selected to do signal conditioning and extract four eigenvalues. The four eigenvalues are the key features for training the artificial neural network analysis model that gives an efficiency prediction algorithm of blood glucose level. The experiment shows that the accuracy of the noninvasive method that has the approximate regression value is 0.9534. The noninvasive blood glucose monitoring by means of Near infrared sensors causes less pain and lower risk of infection when compared with the invasive method.