This study was conducted to see the trend of diseases caused by unhealthy lifestyles on
disadvantaged communities and coastal villages around Gresik and Tuban using hybrid
algorithms through the construction of the structure of Bayesian Network. The problem to be
solved in this study is no system that can detect a relationship between unhealthy behavior that
caused the disease. Model of this structure has never been applied directly in the field to detect a
causal events for example, if a behavior is unhealthy will arise disease. Application of this model
needs to be done with a field study to determine and prove the actual benefits of the concept of a
hybrid construction of Bayesian network structure. The purpose of this research is to produce a
software model capable of early detection of disease risk propensity underdeveloped rural and
coastal communities who have unhealthy lifestyles in the form of construction of the structure and
generates a probability value with a tendency disease. The comparison between the structure of
the origin of the structure of the trial results indicate the level of suitability for complete test data
difference of 10% to the original structure, and suitability for the test data is incomplete for more
than 20% depending on the amount of his missing value. The validity of that smoke will have the
tendency has tuberculosis disease, bronchitis or Lung Cancer through the test system is 80% to
90%.