In current study, a bioinformatic‐based network pharmacology was used to identify the osteosarcoma (
OGS
)‐pathological targets and formononetin (
FN
)‐treated targets before the main core predictive biotargets were screened. In addition, all core targets were selected through a number of bioinformatic databases, followed by identification of predominant biological processes and signalling pathways of
FN
anti‐
OGS
. Further, top three core targets of
FN
anti‐
OGS
were determined as oestrogen receptor 1 (
ESR
1), tumour protein p53 (
TP
53), receptor tyrosine‐protein kinase erbB‐2 (
ERBB
2) respectively. In clinical biochemical data, the plasma samples of
OGS
showed the increased trends of alkaline phosphatase, triglyceride, blood glucose, lactate dehydrogenase, high‐sensitive C‐reactive protein and some immune cell counts when referenced to medical criteria. In clinicopathological examination, histological
OGS
sections resulted in increased positive cell counts of neoplastic
ESR
1,
TP
53,
ERBB
2. To further validate these corn proteins in experimental study in vivo,
FN
‐treated tumour‐bearing nude mice showed intracellular reductions of
ESR
1,
TP
53,
ERBB
2 positive expressions, accompanied with visibly reduced tumour weights. Collectively, our bioinformatic and experimental findings disclosed main core targets, biological processes and signalling pathways of
FN
anti‐
OGS
. Interestingly, the top core targets were representatively validated following
FN
treatment in vivo. Therefore, we reasoned that these predictive targets might be the potential biomarkers for screening and treating osteosarcoma.