PENERAPAN MOVING AVERAGE PADA PREDIKSI PENJUALAN ACCU

The problem faced by TIO ACCU is the difficulty of providing stock according to consumer needed, it caused every month has difference sale of product.  The purpose of this research is to create a battery sales prediction system using the Moving Average method. The Moving Average algorithm used for past sales data doesn’t has seasonal trends or elements. This method is applied to predict the number of battery sales in future periods. The results from 5 periods and validity test with 31 sales data for the MAD method is 3.90, for the MSE method is 20.09, and for the MAPE method is 7.72%. Meanwhile, the results of calculation 7 periods with validity test of 29 sales data for the MAD method is 3.70, the MSE method is 18.90, and for the MAPE method is 7.28%. The test results of the battery sales prediction system using the Moving Average method have run well and optimally with accuracy rate of 92.28% for 5-period predictions and 92.72% for 7-period predictions can be classified as very good criteria, because it has an error rate of less than 10%.