The management practices relevant to bulk tank milk quality were studied in 29 dairy farms of Fars Province, Iran. Farm management practices were obtained by completion of a questionnaire and direct observation. Bulk milk was evaluated by performing standard plate count (SPC), preliminary incubation count (PIC), laboratory pasteurization count (LPC), coliform count (CC), somatic cell count (SCC) and detection of the contagious mastitis agents. The farms were divided into low and high SPC groups (below and above 100,000 CFU/mL) based on Iranian standards. Comparisons of the laboratory results between groups were done using two independent samples t-test. The relationships between the laboratory results were studied by Pearson’s correlation coefficients, all after logarithmic transformation. Associations of managerial risk factors (obtained by the questionnaire and one time of observation) with laboratory results were investigated using two independent samples t-test. P-values <0.05 were considered as significant. Both low and high SPC farms had PIC, LPC, CC and SCC levels above the relevant intervention limits, although the low SPC group had lower PIC and CC levels (P<0.001) and numerically lower LPC and SCC levels (P>0.05). Strong correlations were detected between SPC and PIC, SPC and CC, and PIC and CC, but many of the well explained risk factors of undesired milk quality lacked any relation with high bacterial counts of raw milk. This could be due to the small number of the studied farms, almost similar faults in the farms, wrong answers of the employees to the questions and modification of the milking practices in the presence of an inspector. Infections with Staphylococcus aureus and Mycoplasma bovis could be potential problems in the studied farms, contributing to the elevation of SCC and/or SPC levels. Veterinary interventions could not be based on the questionnaire results. Direct and frequent observations of farm routines could be recommended.