In this study, an ordered non-overlapping block bootstrap procedure has been proposed to obtain multi-step forecast regions for unrestricted vector autoregressive models. The proposed method is not based on either backward or forward representations, so it can be implemented to VARMA or VAR-GARCH models. Also, it is computationally more efficient than the existing techniques. Its finite sample performance is investigated by Monte Carlo experiments and two-real world examples. Our findings show that the proposed method is a good alternative to the available resampling methods and produces better results for long-term forecasting when the model is near non-stationary or near-cointegrated.