Purpose: To investigate the effect of input parameters for an inverse optimization algorithm, and dosimetrically evaluate and compare clinical treatment plans made by inverse and forward planning in high-dose-rate interstitial breast implants. Material and methods: By using a representative breast implant, input parameters responsible for target coverage and dose homogeneity were changed step-by-step, and their optimal values were determined. Then, effects of parameters on dosimetry of normal tissue and organs at risk were investigated. The role of dwell time modulation restriction was also studied. With optimal input parameters, treatment plans of forty-two patients were recalculated using an inverse optimization algorithm (HIPO). Then, a pair-wise comparison between forward and inverse plans was performed using dose-volume parameters. Results: To find a compromise between target coverage and dose homogeneity, we recommend using weight factors in the range of 70-90 for minimum dose, and in the range of 10-30 for maximum dose. Maximum dose value of 120% with a weight factor of 5 is recommended for normal tissue. Dose constraints for organs at risk did not play an important role, and the dwell time gradient restriction had only minor effect on target dosimetry. In clinical treatment plans, at identical target coverage, the inverse planning significantly increased the dose conformality (COIN, 0.75 vs. 0.69, p < 0.0001) and improved the homogeneity (DNR, 0.35 vs. 0.39, p = 0.0027), as compared to forward planning. All dosimetric parameters for non-target breast, ipsilateral lung, ribs, and heart were significantly better with inverse planning. The most exposed small volumes for skin were less in HIPO plans, but without statistical significance. Volume irradiated by 5% was 173.5 cm 3 in forward and 167.7 cm 3 in inverse plans (p = 0.0247). Conclusions: By using appropriate input parameters, inverse planning can provide dosimetrically superior dose distributions over forward planning in interstitial breast implants.