This research is conducted to analyse the pattern of the smart home application user's electricity consumption. Data that is used comes from the sensors installed at a smart home prototype. The data is clustered and the result is used to notify the user whether the smart home daily electricity consumption has exceeds its normal usage. The notification and the real time information about the electricity usage (in KWH, Kilo Watt Hour) presented in the app. This research is expected to help the user to take a better action about his home electricity consumption. K-Means method is employed for the clustering as it is able to process data in big numbers efficiently with a short computing time. There are two resulted clusters, i.e. high energy consumption and low energy consumption. The value obtained from the cluster validation using Davies Bouldin Index is 0.417103167 (non-negative and close to 0), so the clustering result is considered optimal.