Abstract: Clean water is one point of sustainable
Development Goals (SDGs), so to keep indicator water
quality must be determine every period. The Discriminant
analysis is an analysis of dependence that is used to
classify objects into several categories. The purpose of
this study were to determine the discriminant model that
consist of dominant factors of water pollution. Samples
were taken from 42 locations in the surrounding area of
Pontianak City. The sample were analyzed in the
laboratory for the contents of ferrum (Fe), dissolved
oxygen (DO), biochemical oxygen demand (BOD).
Width of the river were also considered is an independent
variable. The methodology includes determining the
pollution index that will be used as dependent variable,
testing the assumption of multivariate normality and
similarity of the covariance variance, conducting the
discriminant analysis classification process using the
Apparent Error Rate method. The pollution level of each
location was visualized in a map. The resulting
discriminant model has an accuracy rate of 69%.