New energy (NE) power generation technology has become an important means to solve the problems of global energy demand and environmental governance due to its sustainability and cleanliness. However, restricted by climate and other factors, the instability of its power generation capacity has a serious impact on the voltage multimedia quality evaluation of small microgrids, which even causes damage to important equipment. This paper proposed a multimedia quality evaluation method in the new energy power generation system relying on intelligent robot numerical control technology. According to the state characteristics of different units, the expanded representation models of conventional units, wind turbines, and photovoltaic units were established, respectively. Based on intelligent robot numerical control technology, a multimedia quality monitoring model of the new energy power generation system in random production simulation was established. The planned outage condition of the unit, which was ignored by the current stochastic simulation method, has been fully considered to ensure the rationality of the simulation of intelligent robot numerical control technology. The intelligent robot numerical control technology was used to explore the expected production power change of the unit and to improve the calculation speed and accuracy. In this paper, an appropriate wind and solar power generation model was constructed based on the output characteristics of wind and solar. Secondly, the influence of random faults between units on the subsequent loading rate was studied through probability calculation. The start-stop factors of conventional units were comprehensively analyzed. Then, the expected value of unit start-stop times was obtained. The average relative error of the expected value of the power shortage of the new energy electricity generation system (EGS) was 2.45%, and the error was lower than 2.98%. In the intelligent robot numerical control technology in this paper, the planned outage condition of the unit was considered, which showed the correctness and superiority of the multimedia quality evaluation of the proposed model.