Let B be a centrally symmetric convex polygon of R 2 and p − q be the distance between two points p, q ∈ R 2 in the normed plane whose unit ball is B. For a set T of n points (terminals) in R 2 , a B-network on T is a network N(T ) = (V , E) with the property that its edges are parallel to the directions of B and for every pair of terminals t i and t j , the network N(T ) contains a shortest B-path between them, i.e., a path of length t i − t j . A minimum B-network on T is a B-network of minimum possible length. The problem of finding minimum B-networks has been introduced by Gudmundsson, Levcopoulos, and Narasimhan (APPROX'99) in the case when the unit ball B is a square (and hence the distance p − q is the l 1 or the l ∞ -distance between p and q) and it has been shown recently by Chin, Guo, and Sun (Symposium on Computational Geometry, pp. [393][394][395][396][397][398][399][400][401][402] 2009) to be strongly NP-complete. Several approximation algorithms (with factors 8, 4, 3, and 2) for the minimum Manhattan problem are known. In this paper, we propose a factor 2.5 approximation algorithm for the minimum B-network problem. The algorithm employs a simplified version of the strip-staircase decomposition proposed in our paper (Chepoi et al. in Theor. Comput. Sci. 390:56-69, 2008, and APPROX-RANDOM, pp. 40-51, 2005) and subsequently used in other factor 2 approximation algorithms for the minimum Manhattan problem. N. Catusse · V. Chepoi ( ) · K. Nouioua · Y. Vaxès