The Global Ionospheric TEC Map (GIM) is becoming one of the essential data sources for monitoring the ionospheric space weather. The development of multi-GNSS, including American GPS, Chinese BDS, European Galileo and Russia GLONASS, can provide very abundant measurements for the GIM calculation. The superiority of multi-GNSS for ionospheric research has been demonstrated in the last two decades, such as the rapid and final product of GIM that is generated using the data from more than 300 global stations in the frame of International GNSS service (IGS). Currently, the real-time data stream from about 110 global distributed multi-GNSS stations can be accessed in the frame of International GNSS service (IGS). Compared with the requirement of GIM calculation, however, the distribution of real-time data stream is still very spare and uneven. This results that the performance of real-time GIM is generally poorer than that of final and rapid product. In view this, a new approach for generating the real-time GIM using the current spare and uneven distributed global stations is proposed by combing the real and predicted ionospheric TEC to avoid the limitation of global real-time data stream. The predicted TEC is calculated from a 2-day forecast SH-based ionospheric model and considered as the pseudo- observation at each grid-point (1 degree plus 1 degree in latitude and longitude) and the weight is adjusted following the distances between the corresponding grid-point and nearby stations. The performance of this proposed approach has been validated by the real-time data stream from about 90 global station at the IGS ionospheric associated analysis center CAS (Chinese Academy of Sciences) in the last six months. The IGS final GIM and the ionospheric TEC observation directly from GNSS station is introduced for comparison in the validation, as well as the rate of ionospheric TEC. The preliminary result indicates that the accuracy of real-time GIM is currently about 2.0-3.0 TECu, which is better that the predicted and rapid GIM product, and almost has the similar accuracy with the final GIM product.