Background & objective: Mineral contaminations are the most important environmental problems in the urban area. The objectives of the present study was determination of concentration, spatial distribution and source determination of Iron, Zinc, Calcium, Copper, Manganese, and Nickel by Artificial Neural Network in Sanandaj city.
Materials & Methods: The study area were divided into the five classes and the total number of 105 surface soil samples (0-10 cm) were collected. Contamination factor and Nemerow index were developed for estimation of metal concentration. In addition, the correlation and principle component analysis were conducted to find the origin of the considered elements in the soil samples. Moreover, IDW interpolation technique was applied for mapping the spatial distribution of the elements and ANN (multilayer perceptron) was applied to estimate the concentration.   
Results: Copper concentration in industrial area was very high and in the other hand, green and industrial area had the lowest and highest Nemerow index, respectively. Calcium concentration especially was affected by residential and high traffic areas. On the other hand, industrial and high traffic areas have the greatest impact on the Zinc and Copper concentration and Nickel, Iron and Manganese contamination was under combination effects of natural and anthropogenic factors. Correlation coefficients derived by ANN were 0.821, 0.661, 0.711, 0.767, 0.712, and 0.701 for Copper, Nickel, Zinc, Manganese, Iron, and Calcium, respectively.
Conclusion: The concentration of Zinc, Calcium and Copper are mostly affected by human activities and Nickel, Iron, Manganese concentration are controlled by both anthropogenic and natural sources. According to the results, three-layer neural network is recommended for estimating metal concentration in the soil samples.