Demand forecast has an important role in inventory planning, for a trade business company it is necessary to make a good stock plan so the company can avoid both lost opportunity cost and loss because of product over supplied. This research aimed to improve the demand forecasting of Roti Panggang Padimas in H&D store by implementing intervention approach forecast which before H&D store only rely on judgement forecast to predict the demand of Roti Panggang Padimas.
A set of historical demand data of Roti Panggang Padimas in H&D store was collected then separated into training dataset and testing dataset; graphs was plotted to identify irregular occasions. With the help of the owner of H&D store, the irregular occasions removed by adding interventions to the historical demand data. To perform the quantitative methods, author being helped by WINQSB software which used 12 time series forecast methods. The MAPE value of each method then calculated. The results showed that MAPE value before the intervention were 17,19%, 21,11%, and 11,17% for Chocolate, Milk Vanilla and Mix respectively. After the intervention applied, the MAPE value for each variant was improved to 6,38%, 10,32%, and 1,96% for Chocolate, Milk Vanilla and Mix variants respectively. The error improvement percentages are 62,89% for Chocolate variant, 51,11% for Milk Vanilla variant and 82,45% for Mix variant.