Society is living in a time of strong technological convergence, where new technologies are discovered and extinguished in an ever shorter time frame. This technological revolution has also arrived for the energy distribution infrastructure with the Smart Grid, in which the smart meter being the main equipment of this revolution. Despite the evolution of the Brazilian meter park, the commercial regulation did not go along with this modernization and continues with reference to the conventional metering system and not considering the opportunities to use consumption data comes from out of the billing date, in cases of data transmission failures, being thus possible use statistical tools for billing estimation. In this context, this dissertation aims to evaluate the regulatory rules of billing considering the absences of readings, proposing a new methodology to define how to estimate the billing in the absence of readings, considering previous consumption and using statistical tools to define the value to be billed. The methodology can be divided into two phases: (i) imputation of missing data in the database, resulting from possible transmission errors of the meters; (ii) forecast of electricity consumption per customer. The present work fulfilled the objectives proposed and presented a promising alternative for billing with smart meters using low cost communication technologies that could have low measurement effectiveness, in this case, 100 percent.