Purpose: To compare the outcome of site preparation for zygomatic oncology implants using conventional preparation with rotary drills or piezoelectric surgery with dedicated inserts for placing two zygomatic implants per zygoma according to a split-mouth design. Materials and Methods: Twenty edentulous patients with severely atrophic maxillas not having sufficient bone volume for placing dental implants and less than 4 mm of bone height subantrally had their hemi-maxillas randomised according to a split-mouth design into implant site preparation with conventional rotational drills or piezoelectric surgery. Two zygomatic oncology implants (unthreaded coronal portion) were placed in each hemi-maxilla. Implants that achieved an insertion torque superior to 40 Ncm were immediately loaded with screw-retained metal reinforced acrylic provisional prostheses. Outcome measures were: prosthesis and implant failures, any complications, time to place the implants, presence of post-operative haematoma, and patient's preference by independent assessors. All patients were followed up to 1 year after loading. Results: In two patients drills had also to be used at the piezoelectric surgery side to enable implant sites to be prepared. One implant for the conventional drill group did not achieve an insertion torque superior to 40 Ncm since it fractured the zygoma. No patients dropped out and two distal oncology implants failed in the same patient (one per group), who was not prosthetically rehabilitated. Six complications occurred at drilled sites and three at piezoelectric surgery sites (two patients had bilateral complications), the difference being not statistically significant (P (McNemar's test) = 0.375; odds ratio = 4.00; 95% CI of odds ratio: 0.45 to 35.79). Implant placement with convention drills took on average 14.35 ± 1.76 min and with piezoelectric surgery 23.50 ± 2.26 min, implant placement time being significantly shorter with conventional drilling (difference = 9.15 ± 1.69 min; 95%CI: 8.36 to 9.94 min; P < 0.001). Post-operative haematomas were more frequent at drilled sites (P = 0.001), and 16 patients found both techniques equally acceptable, while four preferred piezoelectric surgery (P = 0.125). Conclusions: Both drilling techniques achieved similar clinical results, but conventional drilling required 9 min less and could be used in all instances, although it was more aggressive. These results may be system-dependent, therefore they cannot be generalised to other zygomatic systems with confidence.