Space-time (ST) processing in navigation receivers is attractive technology because it provides a sufficient number of degrees of freedom for cancelling a large number of wideband and narrowband jammers simultaneously. Low complexity reduced-rank receiver structures are particularly appealing because the number of filter coefficients can be much larger than the rank of the ST signal model. This paper proposes a new Householder-based multistage Wiener filter (HMSWF) structure for estimating the coefficients of an ST navigation receiver. Both floating point and fixed point arithmetic implementations are considered. The computational complexity of the HMSWF compares favorably with the other MSWF methods reported in literature. Furthermore, the use of Householder transformations at each stage ensures unitary blocking and numerically robust behavior even in finite-precision implementations. Simulations studies verify that the proposed ST navigation receiver based on the HMSWF outperforms other MSWF methods in the antijamming (A,J) task.