Infections due to Haemophilus influnezae require prompt treatment using beta-lactam antibiotics. We used a collection of 81 isolates obtained between 1940 and 2001 from several countries. Whole genome sequencing showed the high heterogeneity of these isolates but allowed us to track the acquisition of beta-lactamase, which was first detected in 1980. Modifications of the ftsI gene encoding the penicillin-binding protein 3, PBP3, also involved in resistance to beta-lactams, appeared in 1991. These modifications (G490E, A502V, R517H, and N526K) were associated with resistance to amoxicillin that was not relieved by a beta-lactamase inhibitor (clavulanic acid), but the isolates retained susceptibility to third-generation cephalosporins (3GC). The modeling of the PBP3 structure suggested that these modifications may reduce the accessibility to the PBP3 active site. Other modifications appeared in 1998 and were associated with resistance to 3GC (S357N, M377I, S385T, and L389F). Modeling of the PBP3 structure suggested that they lie near the S379xN motif of the active site of PBP3. Overall resistance to amoxicillin was detected among 25 isolates (30.8%) of this collection. Resistance to sulfonamides was predicted by a genomic approach from the sequences of the folP gene (encoding the dihydropteroate synthase) due to difficulties in interpreting phenotypic anti-microbial testing and found in 13 isolates (16.0%). Our data suggest a slower spread of resistance to sulfonamides, which may be used for the treatment of H. influnezae infections. Genomic analysis may help in the prediction of antibiotic resistance, inform structure-function analysis, and guide the optimal use of antibiotics.