Disclosed is a method for managing obesity, diabetes, hair loss, aging, cardiovascular, or other blood glucose-spike-induced diseases by reducing the post-prandial blood glucose spike, or the glucose shock. The blood glucose spike, or the glucose shock is reduced by generating a person-specific glucose profile for at least one significant meal to tune or train a blood glucose model (kinetic, artificial intelligence or hybrid), and then using the tuned or the trained model embedded in a computation-capable electronic device to compute and recommend a person-specific meal plan and an exercise plan, including semi-continuous meal ingestion and post-meal exercise while sitting at home or office. Advantages over prior art are that the method uses less strenuous exercise with no or less medicine, is person-specific, quantitative and more suitable for use by an individual, a dietician, or a health care practitioner.